← Back to Semester 3 Hub
📚 Subject Code: BCA-305 🏛️ CCSU Meerut University Pattern 📄 3M, 7.5M & 15M Exam Blueprint ⭐ 2017–2022 Recurring University Exam Archive
Exam Scoring Architecture for BCA-305 (Elements of Statistics): The university theory examination carries 75 Marks with a 3-hour duration, structured into Section A (5 questions × 3M = 15M compulsory), Section B (attempt 2 of 3 questions × 7.5M = 15M), and Section C (attempt 3 of 5 questions × 15M = 45M). In this master solved notebook, every recurring question from 2017, 2018, 2019, 2020, 2021, and 2022 is solved with complete step-by-step mathematical calculations, probability derivations, bivariate tables, and custom vector SVG statistical charts.
SECTION A • 3 MARKS EACH

Very Short Answer Questions & Statistical Definitions

CCSU Examination Blueprint: 5 Questions Compulsory • Total 15 Marks (75 Words Limit)

Q1. Define Statistics. Mention three significant applications of statistics in Computer Science and Data Analytics.

🔥 Repeated: 2017, 2018, 2019, 2021, 2022 3 Marks
✍️ Comprehensive Solution:

Statistics is the mathematical science concerned with the collection, organization, presentation, analysis, and interpretation of numerical data to make rational decisions under uncertainty.

Three Key Applications in Computer Science:

  • Machine Learning & AI: Statistical regression, classification algorithms, and Bayes' theorem form the algorithmic engine of predictive models.
  • Algorithm Performance Analysis: Evaluating average-case time/space complexity and stochastic randomized algorithms via variance and expected values.
  • Network Traffic Engineering & Queuing: Poisson distributions model incoming packet arrivals in routers, preventing buffer overflows.

Q2. Differentiate between Population (Universe) and Sample in statistical data analysis.

🔥 Repeated: 2017, 2018, 2020 3 Marks
✍️ Comprehensive Solution:
Comparison Parameter Population (Universe) Sample
Definition The entire aggregate of all possible individuals, items, or observations possessing specified characteristics. A representative subset or fraction selected from the population for investigation.
Statistical Measure Characteristics are termed Parameters (e.g. population mean $\mu$, population variance $\sigma^2$). Characteristics are termed Statistics (e.g. sample mean $\bar{X}$, sample variance $s^2$).
Feasibility Often infinite or prohibitively expensive to measure completely (Census method). Cost-effective, rapid, and feasible for hypothesis testing (Sampling method).

Q3. Differentiate between Qualitative Attributes and Quantitative Variables with examples.

🔥 Repeated: 2017, 2020, 2022 3 Marks
✍️ Comprehensive Solution:
  • Qualitative Attributes: Non-numerical characteristics that describe qualities or categories. They cannot be directly measured on a continuous metric scale, only counted in frequencies (e.g. Gender, Nationality, Eye Color, Marital Status, Educational Level).
  • Quantitative Variables: Measurable characteristics capable of taking distinct numerical values. Classified into:
    • Discrete Variables: Isolated integer counts (e.g. number of bugs in a software build, family size).
    • Continuous Variables: Fractional real values along a continuum (e.g. server execution time in milliseconds, student height, temperature).

Q4. State the Empirical Relationship between Mean, Median, and Mode in a moderately asymmetrical distribution.

🔥 Repeated: 2017, 2018, 2019, 2021, 2022 3 Marks
✍️ Comprehensive Solution:

In a perfectly symmetrical distribution, Mean, Median, and Mode coincide at the center ($\text{Mean} = \text{Median} = \text{Mode}$). However, in a moderately skewed unimodal frequency distribution, Karl Pearson observed the following Empirical Formula:

$$\text{Mode} = 3(\text{Median}) - 2(\text{Mean})$$ Alternative Expressions: $$\text{Mean} - \text{Mode} = 3(\text{Mean} - \text{Median}) \quad \iff \quad \text{Median} = \frac{\text{Mode} + 2\text{Mean}}{3}$$

Q5. What is the Harmonic Mean (HM)? State the relationship between Arithmetic Mean, Geometric Mean, and Harmonic Mean.

🔥 Repeated: 2018, 2019, 2021 3 Marks
✍️ Comprehensive Solution:

The Harmonic Mean (HM) of a set of positive values is the reciprocal of the arithmetic mean of the reciprocals of the individual observations:

$$HM = \frac{N}{\sum \left(\frac{1}{x_i}\right)}$$

Ideal Use Case: Computing average rates, speeds (km/h), and velocity when the distance traveled is constant.

Fundamental Inequality Relation: For any positive set of observations: $$AM \ge GM \ge HM \quad \text{and} \quad GM = \sqrt{AM \times HM}$$ The equality holds strictly if and only if all individual observations are identical ($x_1 = x_2 = \dots = x_n$).

Q6. What is the Coefficient of Variation (C.V.)? State its formula and significance in comparing data consistency.

🔥 Repeated: 2017, 2018, 2020, 2021 3 Marks
✍️ Comprehensive Solution:

The Coefficient of Variation (C.V.), developed by Karl Pearson, is the relative measure of dispersion expressed as a percentage of the arithmetic mean:

$$CV = \frac{\sigma}{\bar{X}} \times 100$$ where $\sigma$ is the Standard Deviation and $\bar{X}$ is the Arithmetic Mean.

Managerial Decision Rule:

  • A series with a lower C.V. is considered more consistent, uniform, stable, and homogenous.
  • A series with a higher C.V. exhibits greater variability, fluctuation, and inconsistency.

Q7. What is Quartile Deviation (Semi-Interquartile Range)? State its formula and coefficient.

🔥 Repeated: 2019, 2021, 2022 3 Marks
✍️ Comprehensive Solution:

The Quartile Deviation (QD) (or Semi-Interquartile Range) is half the mathematical difference between the Upper Quartile ($Q_3$, 75th percentile) and the Lower Quartile ($Q_1$, 25th percentile):

$$QD = \frac{Q_3 - Q_1}{2}$$

Coefficient of Quartile Deviation: The corresponding relative dimensionless measure of dispersion:

$$\text{Coefficient of } QD = \frac{Q_3 - Q_1}{Q_3 + Q_1}$$

Advantage: Highly robust against extreme outlier observations because it depends strictly on the middle 50% of the ranked distribution.

Q8. Why is Standard Deviation ($\sigma$) considered the absolute best measure of dispersion?

🔥 Repeated: 2018, 2020, 2021 3 Marks
✍️ Comprehensive Solution:

Standard Deviation ($\sigma = \sqrt{\frac{\sum (X - \bar{X})^2}{N}}$) is considered the gold standard measure of dispersion because:

  1. Rigidly Defined: It is mathematically exact and utilizes every single observation in the dataset.
  2. Algebraically Sound: Unlike Mean Deviation, it avoids the unscientific step of dropping negative algebraic signs by squaring deviations.
  3. Amenable to Advanced Mathematics: It enables calculation of combined standard deviation across multiple groups, standard errors, analysis of variance (ANOVA), and confidence intervals in inferential statistics.
  4. Least Fluctuations of Sampling: It exhibits the highest sampling stability among all dispersion metrics.

Q9. State the mathematical limits of Karl Pearson's correlation coefficient. What does $r = 0$ signify?

🔥 Repeated: 2017, 2019, 2021, 2022 3 Marks
✍️ Comprehensive Solution:

Karl Pearson's product-moment correlation coefficient ($r$) is strictly bounded between $-1$ and $+1$ inclusive:

$$-1 \le r \le +1$$
  • $r = +1$: Perfect positive linear correlation.
  • $r = -1$: Perfect negative linear correlation.
  • $r = 0$: Indicates absence of any linear association between the two variables. (However, it does not imply complete independence, as strong non-linear curvilinear relationships can still exist while yielding $r = 0$).

Q10. At what coordinates do the two linear regression lines ($Y$ on $X$ and $X$ on $Y$) intersect?

🔥 Repeated: 2018, 2020, 2022 3 Marks
✍️ Comprehensive Solution:

The two regression lines are given by:

$$Y - \bar{Y} = b_{yx}(X - \bar{X}) \quad \text{and} \quad X - \bar{X} = b_{xy}(Y - \bar{Y})$$

If we substitute $X = \bar{X}$ and $Y = \bar{Y}$ into both equations simultaneously, both sides evaluate to $0 = 0$.

Conclusion: The two linear regression lines always intersect at the Point of Means $(\bar{X}, \bar{Y})$. Solving the two regression equations simultaneously as a system of linear equations yields the exact arithmetic means of the bivariate series.

Q11. State the relationship between correlation coefficient $r$ and the two regression coefficients $b_{yx}$ and $b_{xy}$.

🔥 Repeated: 2017, 2018, 2020, 2021 3 Marks
✍️ Comprehensive Solution:

The correlation coefficient $r$ is the geometric mean of the two regression coefficients ($b_{yx}$ and $b_{xy}$):

$$r = \pm\sqrt{b_{yx} \cdot b_{xy}}$$

Three Inviolable Rules:

  1. The algebraic sign of $r$ is identical to the signs of both regression coefficients (all three must be positive, or all three must be negative).
  2. If one regression coefficient is greater than unity ($> 1$), the other must be less than unity ($< 1$).
  3. Their arithmetic mean is greater than or equal to $r$: $\frac{b_{yx} + b_{xy}}{2} \ge r$.

Q12. Define Mutually Exclusive Events and Independent Events in probability theory.

🔥 Repeated: 2017, 2019, 2021 3 Marks
✍️ Comprehensive Solution:
  • Mutually Exclusive (Disjoint) Events: Two events $A$ and $B$ cannot occur simultaneously in a single trial of an experiment. Their intersection is null: $$A \cap B = \emptyset \implies P(A \cap B) = 0$$ Example: Tossing a coin; getting a Head and getting a Tail in the same toss.
  • Independent Events: The occurrence or non-occurrence of event $A$ has zero effect on the probability of occurrence of event $B$: $$P(B|A) = P(B) \implies P(A \cap B) = P(A) \cdot P(B)$$ Example: Tossing a coin and rolling a die simultaneously.

Q13. State the three Kolmogorov Axioms of Probability.

🔥 Repeated: 2019, 2020, 2022 3 Marks
✍️ Comprehensive Solution:

Let $S$ be the sample space and $E$ be any event in $S$. Andrey Kolmogorov established probability on three mathematical axioms:

  1. Axiom 1 (Non-negativity): For any event $A$, the probability is a non-negative real number: $$P(A) \ge 0$$
  2. Axiom 2 (Certainty / Normalization): The probability of the entire sample space $S$ (sure event) equals unity: $$P(S) = 1$$
  3. Axiom 3 (Additivity): If $A$ and $B$ are mutually exclusive events ($A \cap B = \emptyset$), then: $$P(A \cup B) = P(A) + P(B)$$

Q14. State the Probability Mass Function (PMF) of the Binomial Distribution with its parameters, mean, and variance.

🔥 Repeated: 2018, 2020, 2021, 2022 3 Marks
✍️ Comprehensive Solution:

The Binomial Distribution models the number of successes ($r$) in $n$ independent Bernoulli trials, each with constant success probability $p$ and failure probability $q = 1 - p$:

$$P(X = r) = \binom{n}{r} p^r q^{n-r} = \frac{n!}{r!(n - r)!} p^r q^{n-r}, \quad r = 0, 1, 2, \dots, n$$
  • Parameters: $n$ (number of trials) and $p$ (probability of success).
  • Mean ($\mu$): $\mu = np$
  • Variance ($\sigma^2$): $\sigma^2 = npq$ (Note: Since $0 < q < 1$, $\text{Variance} < \text{Mean}$).

Q15. What are the parameters, mean, and variance of the Poisson Distribution?

🔥 Repeated: 2018, 2019, 2021 3 Marks
✍️ Comprehensive Solution:

The Poisson Distribution is a limiting form of the Binomial distribution when $n \to \infty$ and $p \to 0$ such that $np = \lambda$ (a finite constant):

$$P(X = r) = \frac{e^{-\lambda} \lambda^r}{r!}, \quad r = 0, 1, 2, \dots, \infty$$
  • Single Parameter: $\lambda$ (the average arrival / occurrence rate).
  • Unique Statistical Property: $$\text{Mean} = \lambda \quad \text{and} \quad \text{Variance} = \lambda \implies \mathbf{\text{Mean} = \text{Variance}}$$

Q16. What percentage of area lies within 1, 2, and 3 standard deviations of the mean in a Normal Curve?

🔥 Repeated: 2017, 2019, 2021, 2022 3 Marks
✍️ Comprehensive Solution:

Under a bell-shaped Gaussian Normal Distribution with mean $\mu$ and standard deviation $\sigma$, the total area equals $1.0$ ($100\%$). The standard Empirical Rule (68-95-99.7) dictates:

  • $\mu \pm 1\sigma$ ($Z = \pm 1$): Encloses exactly 68.27% of total observations.
  • $\mu \pm 2\sigma$ ($Z = \pm 2$): Encloses exactly 95.45% of total observations. (At $Z = \pm 1.96$, exactly 95.00%).
  • $\mu \pm 3\sigma$ ($Z = \pm 3$): Encloses exactly 99.73% of total observations (basis of 3-sigma control charts in SQC).

Q17. What are the Time Reversal Test and Factor Reversal Test for an ideal index number?

🔥 Repeated: 2017, 2019, 2020 3 Marks
✍️ Comprehensive Solution:

Introduced by Irving Fisher to evaluate the scientific consistency of index number formulas:

  • Time Reversal Test: The index number formula should work forward and backward with respect to time such that interchanging base year ($0$) and current year ($1$) yields reciprocal values: $$P_{01} \times P_{10} = 1$$
  • Factor Reversal Test: Interchanging prices ($p$) and quantities ($q$) in the price index formula should yield the true Value Index ($V_{01}$): $$P_{01} \times Q_{01} = \frac{\sum p_1 q_1}{\sum p_0 q_0}$$

Q18. What is Statistical Quality Control (SQC)? Distinguish between Chance and Assignable causes of variation.

🔥 Repeated: 2018, 2019, 2021, 2022 3 Marks
✍️ Comprehensive Solution:

Statistical Quality Control (SQC) is the systematic application of statistical tools (primarily Shewhart control charts) to inspect, regulate, and maintain industrial product quality at minimum production cost.

  • Chance (Random) Causes: Inherent, uncontrollable random background variations present in every process (e.g. slight temperature fluctuations, machine vibration). A process with only chance variation is in Statistical Control.
  • Assignable (Special) Causes: Systematic, identifiable operational faults (e.g. substandard raw materials, worn-out cutting tools, operator fatigue). Can be detected and eliminated.
SECTION B • 7.5 MARKS EACH

Mathematical Proofs, Applied Statistics & Control Methods

CCSU Examination Blueprint: Attempt Any 2 of 3 Questions • Total 15 Marks (~200 Words + Derivations)

Q1. State and prove three fundamental mathematical properties of the Arithmetic Mean.

🔥 Repeated: 2017, 2018, 2020, 2021 7.5 Marks
✍️ Comprehensive Solution:
Property 1: The algebraic sum of deviations of all observations from their arithmetic mean is strictly zero.

Proof: Let $x_1, x_2, \dots, x_N$ be $N$ observations with mean $\bar{X} = \frac{\sum x_i}{N} \implies \sum x_i = N\bar{X}$.

$$\sum_{i=1}^N (x_i - \bar{X}) = \sum x_i - \sum \bar{X} = N\bar{X} - N\bar{X} = 0 \quad \blacksquare$$
Property 2: The sum of squared deviations from the mean is minimum.

For any arbitrary constant $A \ne \bar{X}$, $\sum (x_i - \bar{X})^2 < \sum (x_i - A)^2$.

Proof: Let $S = \sum (x_i - A)^2 = \sum [(x_i - \bar{X}) + (\bar{X} - A)]^2$. Expanding:

$$S = \sum (x_i - \bar{X})^2 + 2(\bar{X} - A)\sum (x_i - \bar{X}) + \sum (\bar{X} - A)^2$$

Since $\sum (x_i - \bar{X}) = 0$, the middle term vanishes:

$$\sum (x_i - A)^2 = \sum (x_i - \bar{X})^2 + N(\bar{X} - A)^2$$

Since $N(\bar{X} - A)^2 > 0$ for any $A \ne \bar{X}$, it follows that $\sum (x_i - \bar{X})^2$ is the absolute minimum! $\blacksquare$

Property 3: Combined Arithmetic Mean of Two Groups

If group 1 has $N_1$ items with mean $\bar{X}_1$, and group 2 has $N_2$ items with mean $\bar{X}_2$, the combined mean $\bar{X}_{12}$ is:

$$\bar{X}_{12} = \frac{N_1\bar{X}_1 + N_2\bar{X}_2}{N_1 + N_2}$$

Q2. Explain the graphic method of locating the Median using "Less-than" and "More-than" Ogives with step-by-step procedures.

🔥 Repeated: 2017, 2019, 2021, 2022 7.5 Marks
✍️ Comprehensive Solution:

An Ogive (Cumulative Frequency Polygon) is a smooth graph representing cumulative frequencies. The Median ($M$) can be determined graphically using two standard approaches:

Method 1: Intersection of Two Ogives (Standard Approach)
  1. Construct "Less-than" Frequency Table: Plot cumulative frequencies against upper class boundaries. Connect coordinates with a smooth curve (rises continuously from bottom-left to top-right).
  2. Construct "More-than" Frequency Table: Plot cumulative frequencies against lower class boundaries. Connect coordinates with a smooth curve (falls continuously from top-left to bottom-right).
  3. Locate Point of Intersection: The two curves cross at a unique intersection point $P$. The vertical Y-coordinate of point $P$ is $\frac{N}{2}$.
  4. Read the Median: Draw a vertical perpendicular from intersection point $P$ down to the horizontal X-axis. The point of impact on the X-axis is the exact Median ($M$).
Method 2 (Single Less-Than Ogive): Locate $\frac{N}{2}$ on the vertical Y-axis, draw a horizontal line to intersect the less-than curve, and drop a perpendicular down to the X-axis to read the Median.

Q3. Differentiate between Mean Deviation and Standard Deviation. Why does Mean Deviation ignore algebraic signs, and why is Standard Deviation mathematically superior?

🔥 Repeated: 2018, 2020, 2021 7.5 Marks
✍️ Comprehensive Solution:
Comparison Parameter Mean Deviation (MD) Standard Deviation ($\sigma$)
Formula $$MD = \frac{\sum |X - \bar{X}|}{N}$$ $$\sigma = \sqrt{\frac{\sum (X - \bar{X})^2}{N}}$$
Treatment of Signs Artificially ignores negative signs by taking absolute values ($|d|$). Squares deviations ($(X - \bar{X})^2$), eliminating negative signs algebraically without violating math rules.
Base Point Can be calculated from Mean, Median, or Mode (minimum about Median). Exclusively calculated about the Arithmetic Mean.
Mathematical Utility Cannot be used in advanced algebraic manipulations; cannot calculate combined MD. Highly amenable to algebraic manipulation (Combined SD, standard error, regression, ANOVA).
Sampling Stability Subject to noticeable sampling fluctuations. Least affected by sampling fluctuations; most stable and reliable.

Q4. State and prove the Addition Theorem of Probability for any two non-mutually exclusive events $A$ and $B$.

🔥 Repeated: 2017, 2018, 2020, 2022 7.5 Marks
✍️ Comprehensive Solution:

Theorem Statement: If $A$ and $B$ are any two events associated with a random experiment, then the probability of occurrence of at least one of the events is given by:

$$P(A \cup B) = P(A) + P(B) - P(A \cap B)$$
Mathematical Proof:

From set theory, the compound event $A \cup B$ can be decomposed into two mutually exclusive disjoint sets:

$$A \cup B = A \cup (B - A) = A \cup (B \cap A')$$

Applying the third axiom of probability for disjoint sets:

$$P(A \cup B) = P(A) + P(B \cap A') \quad \dots \text{(Equation 1)}$$

Similarly, the set $B$ can be expressed as the union of two mutually disjoint sets:

$$B = (B \cap A) \cup (B \cap A')$$

Taking probabilities on both sides:

$$P(B) = P(A \cap B) + P(B \cap A') \implies P(B \cap A') = P(B) - P(A \cap B) \quad \dots \text{(Equation 2)}$$

Substituting Equation (2) into Equation (1) yields:

$$P(A \cup B) = P(A) + P(B) - P(A \cap B) \quad \blacksquare$$

Corollary: If $A$ and $B$ are mutually exclusive ($A \cap B = \emptyset \implies P(A \cap B) = 0$), then $P(A \cup B) = P(A) + P(B)$.

Q5. Define Conditional Probability. State and prove the Multiplication Theorem of Probability for dependent and independent events.

🔥 Repeated: 2018, 2019, 2021 7.5 Marks
✍️ Comprehensive Solution:

Conditional Probability: The probability of occurrence of event $B$, given that event $A$ has already occurred ($P(A) > 0$), is denoted by $P(B|A)$ and defined as:

$$P(B|A) = \frac{P(A \cap B)}{P(A)}$$
Multiplication Theorem Statement & Proof:

The probability of the simultaneous occurrence of two events $A$ and $B$ is the product of the probability of $A$ and the conditional probability of $B$ given $A$:

$$P(A \cap B) = P(A) \cdot P(B|A) = P(B) \cdot P(A|B)$$

Proof: Let $S$ have $n$ equally likely outcomes, with $n(A)$ outcomes favorable to $A$, and $n(A \cap B)$ favorable to both $A$ and $B$.

$$P(A \cap B) = \frac{n(A \cap B)}{n} = \frac{n(A)}{n} \times \frac{n(A \cap B)}{n(A)} = P(A) \cdot P(B|A) \quad \blacksquare$$

Extension for Independent Events: If $A$ and $B$ are statistically independent, the occurrence of $A$ does not influence $B$, so $P(B|A) = P(B)$. Thus:

$$P(A \cap B) = P(A) \cdot P(B)$$

Q6. State and prove that the Arithmetic Mean of the two regression coefficients is greater than or equal to the correlation coefficient $r$.

🔥 Repeated: 2017, 2020, 2021, 2022 7.5 Marks
✍️ Comprehensive Solution:

Theorem Statement: If $b_{yx}$ and $b_{xy}$ are the two regression coefficients and $r$ is the correlation coefficient, then:

$$\frac{b_{yx} + b_{xy}}{2} \ge r$$
Mathematical Proof:

For any two real numbers $a$ and $b$, the square of their difference is always non-negative:

$$(a - b)^2 \ge 0 \implies a^2 + b^2 - 2ab \ge 0 \implies a^2 + b^2 \ge 2ab \implies \frac{a^2 + b^2}{2} \ge ab$$

Let $a = \sqrt{b_{yx}}$ and $b = \sqrt{b_{xy}}$. Substituting these into the inequality:

$$\frac{(\sqrt{b_{yx}})^2 + (\sqrt{b_{xy}})^2}{2} \ge \sqrt{b_{yx}} \cdot \sqrt{b_{xy}}$$ $$\frac{b_{yx} + b_{xy}}{2} \ge \sqrt{b_{yx} \cdot b_{xy}}$$

Since the geometric mean of the regression coefficients equals the correlation coefficient ($r = \sqrt{b_{yx} \cdot b_{xy}}$):

$$\frac{b_{yx} + b_{xy}}{2} \ge r \quad \blacksquare$$

The equality holds strictly if and only if $b_{yx} = b_{xy} = r$, which occurs when the two standard deviations are identical ($\sigma_x = \sigma_y$).

Q7. Explain Spearman's Rank Correlation Coefficient formula. How are tied / repeated ranks adjusted using the correction factor?

🔥 Repeated: 2017, 2019, 2021 7.5 Marks
✍️ Comprehensive Solution:

Charles Spearman's Rank Correlation Coefficient ($r_s$) measures the strength of association between two ranked qualitative or quantitative variables.

1. Standard Formula (No Tied Ranks):
$$r_s = 1 - \frac{6 \sum D_i^2}{n(n^2 - 1)}$$

where $D_i = R_{1i} - R_{2i}$ is the difference between ranks assigned to the $i$-th individual, and $n$ is total pairs.

2. Adjustment for Tied / Repeated Ranks:

When two or more items have identical values, each is assigned the average rank (e.g. if two items tie for 3rd and 4th position, each receives rank $\frac{3 + 4}{2} = 3.5$).

To compensate for the loss of variance caused by tie-averaging, a Correction Factor ($CF$) is added to $\sum D^2$ for each group of tied items:

$$CF = \frac{m(m^2 - 1)}{12}$$

where $m$ is the number of times an item is repeated. The modified formula becomes:

$$r_s = 1 - \frac{6 \left[ \sum D^2 + \sum \frac{m_i(m_i^2 - 1)}{12} \right]}{n(n^2 - 1)}$$

Q8. Prove that for a Binomial distribution with parameters $n$ and $p$, the Mean is $np$ and the Variance is $npq$.

🔥 Repeated: 2018, 2019, 2022 7.5 Marks
✍️ Comprehensive Solution:

Let $X \sim B(n, p)$. The Probability Mass Function is $P(X = r) = \binom{n}{r} p^r q^{n-r}$.

1. Derivation of Mean:
$$E(X) = \sum_{r=0}^n r \cdot \binom{n}{r} p^r q^{n-r} = \sum_{r=1}^n r \cdot \frac{n}{r} \binom{n-1}{r-1} p \cdot p^{r-1} q^{(n-1)-(r-1)}$$ $$E(X) = np \sum_{r=1}^n \binom{n-1}{r-1} p^{r-1} q^{(n-1)-(r-1)} = np (q + p)^{n-1}$$

Since $q + p = 1$, we get: $\mathbf{E(X) = \text{Mean} = np} \quad \blacksquare$

2. Derivation of Variance:

Recall that $E(X(X - 1)) = E(X^2) - E(X)$.

$$E(X(X - 1)) = \sum_{r=0}^n r(r - 1) \binom{n}{r} p^r q^{n-r} = n(n - 1)p^2 \sum_{r=2}^n \binom{n-2}{r-2} p^{r-2} q^{(n-2)-(r-2)} = n(n - 1)p^2$$ $$E(X^2) = E(X(X - 1)) + E(X) = n(n - 1)p^2 + np = n^2 p^2 - np^2 + np$$

Now, by definition of variance:

$$\text{Var}(X) = E(X^2) - [E(X)]^2 = (n^2 p^2 - np^2 + np) - (np)^2 = np(1 - p) = \mathbf{npq} \quad \blacksquare$$

Q9. Explain the construction and interpretation of $\bar{X}$ (Mean) and $R$ (Range) Control Charts in Statistical Quality Control.

🔥 Repeated: 2018, 2019, 2021, 2022 7.5 Marks
✍️ Comprehensive Solution:

In manufacturing quality assurance, $\bar{X}$ and $R$ Charts are paired variable control charts used to monitor processes producing measurable dimensions (weight, diameter, resistance):

  • $\bar{X}$ Chart: Monitors process centering (location/mean shifts).
  • $R$ Chart: Monitors process variability (dispersion/spread). The $R$ chart is always analyzed first because reliable $\bar{X}$ limits require dispersion to be in control.
Control Limits Formulas (for $k$ subgroups of size $n$):

Compute grand mean $\bar{\bar{X}} = \frac{\sum \bar{X}}{k}$ and average range $\bar{R} = \frac{\sum R}{k}$. Using standard statistical factor tables ($A_2, D_3, D_4$):

Chart Type Lower Control Limit (LCL) Central Line (CL) Upper Control Limit (UCL)
$\bar{X}$ Chart $$LCL_{\bar{X}} = \bar{\bar{X}} - A_2 \bar{R}$$ $$CL_{\bar{X}} = \bar{\bar{X}}$$ $$UCL_{\bar{X}} = \bar{\bar{X}} + A_2 \bar{R}$$
$R$ Chart $$LCL_R = D_3 \bar{R} \quad (\text{0 if } n \le 6)$$ $$CL_R = \bar{R}$$ $$UCL_R = D_4 \bar{R}$$

Interpretation: If all sample points oscillate randomly between LCL and UCL without exhibiting runs or trends, the process is deemed in Statistical Quality Control.

Q10. Explain the four components of a Time Series with practical business examples.

🔥 Repeated: 2017, 2018, 2020, 2022 7.5 Marks
✍️ Comprehensive Solution:

A Time Series is an arrangement of statistical observations recorded chronologically over successive time intervals ($Y_t$). Classical decomposition breaks variation into four components:

  • 1. Secular Trend ($T$): The smooth, persistent, long-term general tendency of a series to increase or decrease over decades (e.g. gradual growth of world internet users, secular rise in agricultural production due to mechanized farming).
  • 2. Seasonal Variations ($S$): Rhythmic, periodic fluctuations that repeat systematically within a single calendar year (duration $< 12$ months) caused by climate seasons or cultural festivals (e.g. surge in AC sales in May/June, festive spike in electronic appliance purchases during Diwali).
  • 3. Cyclical Fluctuations ($C$): Wave-like oscillatory patterns spanning multi-year horizons (3 to 10 years) caused by macroeconomic business cycles (Phases: Prosperity / Boom → Recession → Depression → Recovery).
  • 4. Irregular / Random Noise ($I$): Erratic, unpredictable, non-recurring residual shocks caused by unforeseen external catastrophes (e.g. sudden earthquakes, flash floods, war hostilities, pandemic lockdown disruptions).
SECTION C • 15 MARKS EACH

Master Numerical Problems, Mathematical Proofs & SVG Diagrams

CCSU Examination Blueprint: Attempt Any 3 of 5 Questions • Total 45 Marks (~500+ Words + Step-by-Step Numericals)

Q1. Explain the properties of the Normal Distribution curve in detail. Illustrate the 68-95-99.7 Empirical Rule with an architectural SVG diagram and solve a complete numerical on student examination scores.

🔥 Repeated: 2017, 2018, 2019, 2021, 2022 15 Marks
✍️ Comprehensive Solution:
1. Mathematical Equation & Core Properties of the Normal Curve

The Normal (Gaussian) Distribution is a continuous probability distribution governed by the probability density function (PDF):

$$f(x) = \frac{1}{\sigma \sqrt{2\pi}} e^{-\frac{(x - \mu)^2}{2\sigma^2}}, \quad -\infty < x < \infty$$

where $\mu$ is the population mean and $\sigma$ is the standard deviation.

Salient Characteristics:

  • Symmetrical Bell-Shape: Perfectly symmetrical about the vertical ordinate at $X = \mu$. Skewness is zero ($\beta_1 = 0$) and Kurtosis is mesokurtic ($\beta_2 = 3$).
  • Unimodal Coincidence: Mean, Median, and Mode are exactly identical: $\text{Mean} = \text{Median} = \text{Mode} = \mu$.
  • Asymptotic Tails: The curve approaches the horizontal axis indefinitely on both sides ($-\infty$ to $+\infty$) but never touches it.
  • Points of Inflection: The curvature transitions from concave to convex at $X = \mu \pm \sigma$.
  • Standard Normal Variate ($Z$): Transformed via $Z = \frac{X - \mu}{\sigma}$ yielding $Z \sim N(0, 1)$ with mean $0$ and variance $1$. Total area under the curve is strictly $1.0$ ($100\%$).
2. Standard Normal Bell Curve Architecture Diagram
STANDARD NORMAL DISTRIBUTION: 68% - 95% - 99.7% EMPIRICAL RULE
μ (Z = 0) μ - 1σ μ + 1σ μ - 2σ μ + 2σ μ - 3σ μ + 3σ ← μ ± 1σ : 68.27% → ← μ ± 2σ : 95.45% → ← μ ± 3σ : 99.73% of Process Distribution →
3. Step-by-Step Numerical Solution
Problem Statement: The marks obtained by $N = 1000$ university students are normally distributed with Mean $\mu = 60$ and Standard Deviation $\sigma = 10$. Calculate the expected number of students scoring above 75 marks. (Given: Area under standard normal curve between $Z = 0$ and $Z = 1.5$ is $0.4332$).

Step 1: Convert raw score $X = 75$ into standard normal score $Z$:

$$Z = \frac{X - \mu}{\sigma} = \frac{75 - 60}{10} = \frac{15}{10} = +1.5$$

Step 2: Determine tail area probability $P(X > 75) = P(Z > 1.5)$:

By symmetry, the entire area to the right of the mean ($Z = 0$) is $0.5000$.

$$P(Z > 1.5) = 0.5000 - P(0 \le Z \le 1.5) = 0.5000 - 0.4332 = 0.0668$$

Step 3: Multiply by total student population $N = 1000$:

$$\text{Expected Number of Students} = N \times P(Z > 1.5) = 1000 \times 0.0668 = 66.8 \approx \mathbf{67 \text{ Students}}$$

Q2. Explain the construction of "Less-than" and "More-than" Cumulative Frequency Tables and Ogive curves. Diagrammatically prove how their intersection point gives the exact Median ($M$).

🔥 Repeated: 2017, 2019, 2020, 2021 15 Marks
✍️ Comprehensive Solution:
1. Construction Rules for the Two Ogives
Feature "Less-Than" Ogive "More-Than" Ogive
X-Coordinate Points Upper class limits of class intervals. Lower class limits of class intervals.
Y-Coordinate Points Cumulative frequency summed from lowest to highest class. Cumulative frequency summed from highest down to lowest class.
Curve Trajectory Rising S-shape: Starts at $(0, 0)$ and climbs to $(X_{\max}, N)$. Falling S-shape: Starts at $(X_{\min}, N)$ and descends to $(X_{\max}, 0)$.
2. Ogives Intersection Architecture Diagram
GRAPHICAL DETERMINATION OF MEDIAN: OGIVES INTERSECTION
Class Boundaries (Marks) Cumulative Frequency (cf) N (100) Less-Than Ogive More-Than Ogive P (Intersection) N / 2 (50) Median (M)
3. Mathematical Justification

At any observation value $X$, the Less-Than cumulative frequency counts the total observations $\le X$, while the More-Than cumulative frequency counts observations $\ge X$. When the two values equate at intersection point $P$:

$$cf_{\text{less}} = cf_{\text{more}} = \frac{N}{2}$$

By definition, the value dividing a ranked distribution into two exactly equal halves containing 50% of total observations is the Median ($M$). Thus, the perpendicular dropped from intersection point $P$ to the horizontal axis guarantees the exact median value!

Q3. Calculate Arithmetic Mean, Standard Deviation, and Coefficient of Variation for the following grouped frequency distribution. Compare consistency across two production factories.

🔥 Repeated: 2017, 2018, 2020, 2021, 2022 15 Marks
✍️ Comprehensive Solution:
1. Grouped Frequency Calculation Table

Assume class width $h = 10$, Assumed Mean $A = 35$, step deviation $d' = \frac{m - A}{h} = \frac{m - 35}{10}$.

Class Interval Midpoint ($m$) Frequency ($f$) $d' = \frac{m - 35}{10}$ $f \cdot d'$ $f \cdot (d')^2$
10 – 20156-2-1224
20 – 302512-1-1212
30 – 4035 ($A$)18000
40 – 504510+1+1010
50 – 60554+2+816
Total -- $N = \sum f = 50$ -- $\sum f d' = -6$ $\sum f (d')^2 = 62$
2. Step-by-Step Statistical Formulas

A. Arithmetic Mean ($\bar{X}$):

$$\bar{X} = A + \left(\frac{\sum f d'}{N}\right) \times h = 35 + \left(\frac{-6}{50}\right) \times 10 = 35 - 1.2 = \mathbf{33.8}$$

B. Standard Deviation ($\sigma$):

$$\sigma = h \times \sqrt{\frac{\sum f (d')^2}{N} - \left(\frac{\sum f d'}{N}\right)^2} = 10 \times \sqrt{\frac{62}{50} - \left(\frac{-6}{50}\right)^2}$$ $$\sigma = 10 \times \sqrt{1.24 - (-0.12)^2} = 10 \times \sqrt{1.24 - 0.0144} = 10 \times \sqrt{1.2256} = 10 \times 1.107 = \mathbf{11.07}$$

C. Coefficient of Variation ($CV$):

$$CV = \frac{\sigma}{\bar{X}} \times 100 = \frac{11.07}{33.8} \times 100 = \mathbf{32.75\%}$$
3. Factory Consistency Comparison
Exam Benchmark Problem: Two factories $A$ and $B$ report weekly wages:
• Factory A: $N_1 = 50$, Mean $\bar{X}_1 = \text{Rs. } 300$, Variance $\sigma_1^2 = 25 \implies \sigma_1 = 5$.
• Factory B: $N_2 = 100$, Mean $\bar{X}_2 = \text{Rs. } 250$, Variance $\sigma_2^2 = 36 \implies \sigma_2 = 6$.
Analysis:
1. Total Wage Bill: Factory A pays $50 \times 300 = \text{Rs. } 15,000$. Factory B pays $100 \times 250 = \text{Rs. } 25,000$. Hence, Factory B pays a larger total wage bill.
2. Wage Consistency: $CV_A = \frac{5}{300} \times 100 = \mathbf{1.67\%}$, while $CV_B = \frac{6}{250} \times 100 = \mathbf{2.40\%}$. Since $CV_A < CV_B$, Factory A displays significantly greater consistency and uniformity in wages.

Q4. Explain Karl Pearson's Coefficient of Correlation ($r$). Prove mathematically that $-1 \le r \le +1$. Draw comprehensive SVG scatter diagrams illustrating positive, negative, and zero correlation patterns.

🔥 Repeated: 2017, 2018, 2019, 2020, 2021, 2022 15 Marks
✍️ Comprehensive Solution:
1. Definition and Mathematical Formulas

Karl Pearson's Coefficient of Correlation ($r$) measures the degree and direction of linear association between two quantitative continuous variables $X$ and $Y$:

$$r = \frac{\text{Cov}(X, Y)}{\sigma_x \cdot \sigma_y} = \frac{\sum (X - \bar{X})(Y - \bar{Y})}{\sqrt{\sum (X - \bar{X})^2 \cdot \sum (Y - \bar{Y})^2}}$$

Shortcut Machine Formula for Raw Scores:

$$r = \frac{N \sum XY - (\sum X)(\sum Y)}{\sqrt{\left[N \sum X^2 - (\sum X)^2\right] \left[N \sum Y^2 - (\sum Y)^2\right]}}$$
2. Scatter Diagrams Architecture
SCATTER DIAGRAMS: CORRELATION TAXONOMY
A. Strong Positive (r ≈ +0.95) B. Strong Negative (r ≈ -0.90) C. Zero Correlation (r ≈ 0)
3. Mathematical Proof that $-1 \le r \le +1$

Let $x_i = X_i - \bar{X}$ and $y_i = Y_i - \bar{Y}$. Consider the sum of squares of standardized z-scores for any real parameter $k$:

$$\sum \left( \frac{x_i}{\sigma_x} \pm \frac{y_i}{\sigma_y} \right)^2 \ge 0$$

Expanding the algebraic square:

$$\frac{\sum x_i^2}{\sigma_x^2} + \frac{\sum y_i^2}{\sigma_y^2} \pm 2 \frac{\sum x_i y_i}{\sigma_x \sigma_y} \ge 0$$

By definition, $\sum x_i^2 = N \sigma_x^2$, $\sum y_i^2 = N \sigma_y^2$, and $\sum x_i y_i = N \cdot \text{Cov}(X, Y) = N \cdot r \sigma_x \sigma_y$:

$$\frac{N \sigma_x^2}{\sigma_x^2} + \frac{N \sigma_y^2}{\sigma_y^2} \pm 2 \frac{N \cdot r \sigma_x \sigma_y}{\sigma_x \sigma_y} \ge 0$$ $$N + N \pm 2N r \ge 0 \implies 2N(1 \pm r) \ge 0$$

Dividing by $2N > 0$:

$$1 + r \ge 0 \implies r \ge -1 \quad \text{and} \quad 1 - r \ge 0 \implies r \le +1$$ $$\therefore \mathbf{-1 \le r \le +1} \quad \blacksquare$$

Q5. Derive the two linear regression lines ($Y$ on $X$ and $X$ on $Y$). Prove that the acute angle $\theta$ between the two lines satisfies $\tan\theta = \left(\frac{1 - r^2}{|r|}\right) \left(\frac{\sigma_x \sigma_y}{\sigma_x^2 + \sigma_y^2}\right)$, and explain the limiting cases when $r = 0$ and $r = \pm 1$.

🔥 Repeated: 2017, 2019, 2020, 2022 15 Marks
✍️ Comprehensive Solution:
1. Equations and Slopes of the Two Regression Lines
  • Line of Regression of $Y$ on $X$: Minimizes the sum of vertical squared errors; used to estimate $Y$ for a known $X$: $$Y - \bar{Y} = b_{yx}(X - \bar{X}) \quad \text{where slope } m_1 = b_{yx} = r \frac{\sigma_y}{\sigma_x}$$
  • Line of Regression of $X$ on $Y$: Minimizes the sum of horizontal squared errors; used to estimate $X$ for a known $Y$: $$X - \bar{X} = b_{xy}(Y - \bar{Y}) \implies Y - \bar{Y} = \frac{1}{b_{xy}}(X - \bar{X}) \quad \text{where slope } m_2 = \frac{1}{b_{xy}} = \frac{1}{r} \frac{\sigma_y}{\sigma_x}$$
2. Regression Lines & Angle $\theta$ Diagram
INTERSECTING REGRESSION LINES AT POINT OF MEANS (&Xmacr;, &Ymacr;)
Variable X Variable Y Y on X (Slope = byx) X on Y (Slope = 1/bxy) Point of Means (&Xmacr;, &Ymacr;) &Ymacr; &Xmacr; θ
3. Derivation of Acute Angle $\theta$

Using the coordinate geometry angle formula $\tan\theta = \left|\frac{m_2 - m_1}{1 + m_1 m_2}\right|$:

$$\tan\theta = \left|\frac{\frac{1}{r}\frac{\sigma_y}{\sigma_x} - r\frac{\sigma_y}{\sigma_x}}{1 + \left(r\frac{\sigma_y}{\sigma_x}\right)\left(\frac{1}{r}\frac{\sigma_y}{\sigma_x}\right)}\right| = \left|\frac{\frac{\sigma_y}{\sigma_x} \left(\frac{1 - r^2}{r}\right)}{1 + \frac{\sigma_y^2}{\sigma_x^2}}\right| = \left|\frac{\frac{\sigma_y}{\sigma_x}\left(\frac{1 - r^2}{r}\right)}{\frac{\sigma_x^2 + \sigma_y^2}{\sigma_x^2}}\right|$$ $$\tan\theta = \left(\frac{1 - r^2}{|r|}\right) \left(\frac{\sigma_x \sigma_y}{\sigma_x^2 + \sigma_y^2}\right) \quad \blacksquare$$
4. Interpretation of Limiting Cases:
  • Case 1: When $r = \pm 1$ (Perfect Correlation): $$\tan\theta = \left(\frac{1 - (\pm 1)^2}{1}\right) (\dots) = 0 \implies \mathbf{\theta = 0^\circ}$$ The two regression lines coincide into a single unified line.
  • Case 2: When $r = 0$ (Zero Correlation): $$\tan\theta = \frac{1 - 0}{0} = \infty \implies \mathbf{\theta = 90^\circ}$$ The two regression lines are mutually perpendicular to each other (parallel to the coordinate axes).

Q6. State and derive Bayes' Theorem of Inverse Probability. Solve a classic three-machine industrial defect problem step-by-step.

🔥 Repeated: 2017, 2018, 2019, 2021, 2022 15 Marks
✍️ Comprehensive Solution:
1. Statement and Derivation of Bayes' Theorem

Let $E_1, E_2, \dots, E_k$ be $k$ mutually exclusive and exhaustive events (hypotheses) constituting a partition of sample space $S$ with prior probabilities $P(E_i) > 0$. If $A$ is an observed event ($P(A) > 0$), then the posterior probability of hypothesis $E_i$ given that $A$ has occurred is:

$$P(E_i | A) = \frac{P(E_i) \cdot P(A | E_i)}{\sum_{j=1}^k P(E_j) \cdot P(A | E_j)}$$

Proof: By definition of conditional probability:

$$P(E_i | A) = \frac{P(E_i \cap A)}{P(A)} = \frac{P(E_i) \cdot P(A | E_i)}{P(A)} \quad \dots \text{(Equation 1)}$$

By the Law of Total Probability, since $E_1, \dots, E_k$ partition $S$:

$$P(A) = \sum_{j=1}^k P(E_j \cap A) = \sum_{j=1}^k P(E_j) \cdot P(A | E_j) \quad \dots \text{(Equation 2)}$$

Substituting (2) into (1) yields Bayes' theorem. $\blacksquare$

2. Complete Step-by-Step Numerical Solution
Problem Statement: In a manufacturing plant, three machines $M_1, M_2, M_3$ produce respectively 25%, 35%, and 40% of total bolt output. Of their respective outputs, 5%, 4%, and 2% are defective. A bolt is drawn at random and found to be defective. Find the probability that it was manufactured by Machine $M_1$.

Step 1: Define Events and Prior Probabilities:

  • $P(M_1) = 0.25$, $P(M_2) = 0.35$, $P(M_3) = 0.40$
  • Let $D$ be the event of drawing a defective bolt.

Step 2: Conditional Likelihoods $P(D | M_i)$:

  • $P(D | M_1) = 0.05$
  • $P(D | M_2) = 0.04$
  • $P(D | M_3) = 0.02$

Step 3: Total Probability of a Defective Bolt $P(D)$:

$$P(D) = P(M_1)P(D|M_1) + P(M_2)P(D|M_2) + P(M_3)P(D|M_3)$$ $$P(D) = (0.25 \times 0.05) + (0.35 \times 0.04) + (0.40 \times 0.02) = 0.0125 + 0.0140 + 0.0080 = \mathbf{0.0345}$$

Step 4: Compute Posterior Probability $P(M_1 | D)$ via Bayes' Formula:

$$P(M_1 | D) = \frac{P(M_1) \cdot P(D | M_1)}{P(D)} = \frac{0.0125}{0.0345} = \frac{125}{345} = \frac{25}{69} \approx \mathbf{0.3623 \text{ (or } 36.23\%)}$$

Q7. What is Fisher's Ideal Index Number? Why is it considered "Ideal"? Prove that it satisfies both the Time Reversal Test and the Factor Reversal Test, and calculate Fisher's Index for given data.

🔥 Repeated: 2017, 2018, 2019, 2020, 2022 15 Marks
✍️ Comprehensive Solution:
1. Theoretical Foundation: Why Fisher's Index is Called "Ideal"

Fisher's Price Index Number, introduced by American economist Prof. Irving Fisher, is defined as the geometric mean of Laspeyres' Index ($P_{01}^L$) and Paasche's Index ($P_{01}^P$):

$$P_{01}^F = \sqrt{P_{01}^L \times P_{01}^P} = \sqrt{\left(\frac{\sum p_1 q_0}{\sum p_0 q_0} \times 100\right) \times \left(\frac{\sum p_1 q_1}{\sum p_0 q_1} \times 100\right)} = \sqrt{\frac{\sum p_1 q_0}{\sum p_0 q_0} \times \frac{\sum p_1 q_1}{\sum p_0 q_1}} \times 100$$

It is celebrated as an "Ideal" Index Number due to four decisive mathematical and economic attributes:

  1. Geometric Mean Foundation: It utilizes the geometric mean, which is scientifically proven to be the best average for calculating ratios and index numbers.
  2. Symmetrical Weighting: It incorporates both base-year quantities ($q_0$) and current-year quantities ($q_1$), neutralizing the upward bias of Laspeyres' index and the downward bias of Paasche's index.
  3. Satisfies Time Reversal Test (TRT): It works backward as well as forward in time ($P_{01} \times P_{10} = 1$).
  4. Satisfies Factor Reversal Test (FRT): Interchanging price and quantity yields the true Value Index ($P_{01} \times Q_{01} = V_{01} = \frac{\sum p_1 q_1}{\sum p_0 q_0}$).
2. Mathematical Proof of Time Reversal Test (TRT)
Definition of TRT: An index number satisfies the Time Reversal Test if reversing the time periods (exchanging base year $0$ and current year $1$) yields a reciprocal result (omitting the factor of 100): $$P_{01} \times P_{10} = 1$$

Step-by-step Proof:

Forward Index ($0$ to $1$):

$$P_{01}^F = \sqrt{\frac{\sum p_1 q_0}{\sum p_0 q_0} \times \frac{\sum p_1 q_1}{\sum p_0 q_1}}$$

Backward Index ($1$ to $0$), obtained by swapping subscripts $0 \leftrightarrow 1$:

$$P_{10}^F = \sqrt{\frac{\sum p_0 q_1}{\sum p_1 q_1} \times \frac{\sum p_0 q_0}{\sum p_1 q_0}}$$

Multiplying the two expressions:

$$P_{01}^F \times P_{10}^F = \sqrt{\frac{\sum p_1 q_0}{\sum p_0 q_0} \times \frac{\sum p_1 q_1}{\sum p_0 q_1} \times \frac{\sum p_0 q_1}{\sum p_1 q_1} \times \frac{\sum p_0 q_0}{\sum p_1 q_0}}$$

Cancelling identical numerator and denominator terms:

  • $\sum p_1 q_0$ cancels with $\sum p_1 q_0$
  • $\sum p_0 q_0$ cancels with $\sum p_0 q_0$
  • $\sum p_1 q_1$ cancels with $\sum p_1 q_1$
  • $\sum p_0 q_1$ cancels with $\sum p_0 q_1$
$$P_{01}^F \times P_{10}^F = \sqrt{1} = \mathbf{1}$$

Hence, Fisher's Ideal Index satisfies the Time Reversal Test. $\blacksquare$

3. Mathematical Proof of Factor Reversal Test (FRT)
Definition of FRT: An index number satisfies the Factor Reversal Test if interchanging prices and quantities ($p \leftrightarrow q$) produces the true Value Ratio: $$P_{01} \times Q_{01} = \frac{\sum p_1 q_1}{\sum p_0 q_0} = V_{01}$$

Step-by-step Proof:

Fisher's Price Index:

$$P_{01}^F = \sqrt{\frac{\sum p_1 q_0}{\sum p_0 q_0} \times \frac{\sum p_1 q_1}{\sum p_0 q_1}}$$

Fisher's Quantity Index ($Q_{01}^F$), obtained by swapping $p \leftrightarrow q$:

$$Q_{01}^F = \sqrt{\frac{\sum q_1 p_0}{\sum q_0 p_0} \times \frac{\sum q_1 p_1}{\sum q_0 p_1}} = \sqrt{\frac{\sum p_0 q_1}{\sum p_0 q_0} \times \frac{\sum p_1 q_1}{\sum p_1 q_0}}$$

Multiplying price and quantity indices:

$$P_{01}^F \times Q_{01}^F = \sqrt{\frac{\sum p_1 q_0}{\sum p_0 q_0} \times \frac{\sum p_1 q_1}{\sum p_0 q_1} \times \frac{\sum p_0 q_1}{\sum p_0 q_0} \times \frac{\sum p_1 q_1}{\sum p_1 q_0}}$$

Notice that $\sum p_1 q_0$ and $\sum p_0 q_1$ cancel out:

$$P_{01}^F \times Q_{01}^F = \sqrt{\frac{(\sum p_1 q_1)^2}{(\sum p_0 q_0)^2}} = \mathbf{\frac{\sum p_1 q_1}{\sum p_0 q_0} = V_{01}}$$

Hence, Fisher's Ideal Index satisfies the Factor Reversal Test. $\blacksquare$

4. Complete Step-by-Step Numerical Computation
Exam Problem: From the following data, compute Fisher's Ideal Index Number and verify both tests:
Commodity Base Year (2018) Current Year (2022) Working Products
$p_0$ (₹) $q_0$ $p_1$ (₹) $q_1$ $p_1 q_0$ $p_0 q_0$ $p_1 q_1$ $p_0 q_1$
A 6 50 10 56 $10 \times 50 = 500$ $6 \times 50 = 300$ $10 \times 56 = 560$ $6 \times 56 = 336$
B 2 100 2 120 $2 \times 100 = 200$ $2 \times 100 = 200$ $2 \times 120 = 240$ $2 \times 120 = 240$
C 4 60 6 60 $6 \times 60 = 360$ $4 \times 60 = 240$ $6 \times 60 = 360$ $4 \times 60 = 240$
D 10 30 12 24 $12 \times 30 = 360$ $10 \times 30 = 300$ $12 \times 24 = 288$ $10 \times 24 = 240$
Column Totals ($\sum$) $\sum p_1 q_0 = 1420$ $\sum p_0 q_0 = 1040$ $\sum p_1 q_1 = 1448$ $\sum p_0 q_1 = 1056$

Step 1: Compute Component Indices:

Laspeyres' Index:

$$P_{01}^L = \frac{\sum p_1 q_0}{\sum p_0 q_0} \times 100 = \frac{1420}{1040} \times 100 \approx \mathbf{136.54}$$

Paasche's Index:

$$P_{01}^P = \frac{\sum p_1 q_1}{\sum p_0 q_1} \times 100 = \frac{1448}{1056} \times 100 \approx \mathbf{137.12}$$

Step 2: Compute Fisher's Ideal Index:

$$P_{01}^F = \sqrt{P_{01}^L \times P_{01}^P} = \sqrt{136.54 \times 137.12} = \sqrt{\frac{1420}{1040} \times \frac{1448}{1056}} \times 100 = \sqrt{1.36538 \times 1.37121} \times 100$$ $$P_{01}^F = \sqrt{1.87222} \times 100 \approx 1.36829 \times 100 = \mathbf{136.83}$$

Step 3: Verification of Tests with Calculated Figures:

  • Time Reversal Test: $$P_{01} \times P_{10} = \sqrt{\frac{1420}{1040} \times \frac{1448}{1056} \times \frac{1056}{1448} \times \frac{1040}{1420}} = \sqrt{1} = \mathbf{1} \quad \text{(Satisfied)}$$
  • Factor Reversal Test: $$P_{01} \times Q_{01} = \sqrt{\frac{1420}{1040} \times \frac{1448}{1056} \times \frac{1056}{1040} \times \frac{1448}{1420}} = \sqrt{\frac{(1448)^2}{(1040)^2}} = \frac{1448}{1040} = \frac{\sum p_1 q_1}{\sum p_0 q_0} \approx \mathbf{1.3923} \quad \text{(Satisfied)}$$

Q8. Explain the Method of Least Squares for fitting a linear trend equation in Time Series. Derive the normal equations, and fit a straight line trend for 7-year production data to estimate future output.

🔥 Repeated: 2017, 2018, 2020, 2021, 2022 15 Marks
✍️ Comprehensive Solution:
1. Principle of Least Squares & Derivation of Normal Equations

The Method of Least Squares is the most objective, mathematical, and widely used method for estimating the secular trend line. Given time series data $(t_i, Y_i)$ for $i = 1, 2, \dots, N$, we wish to fit a straight line trend:

$$Y = a + bX$$

where $Y$ is the estimated trend value, $X$ is the coded time variable, $a$ is the $Y$-intercept (trend value at origin $X=0$), and $b$ is the slope (rate of annual change in $Y$).

Criterion: The line is fitted such that the sum of the squared vertical deviations between observed values $Y$ and trend values $\hat{Y} = a + bX$ is minimized:

$$S = \sum (Y - \hat{Y})^2 = \sum (Y - a - bX)^2 \to \text{Minimum}$$

Differentiating $S$ partially with respect to $a$ and $b$ and setting to zero:

$$\frac{\partial S}{\partial a} = -2 \sum (Y - a - bX) = 0 \implies \mathbf{\sum Y = Na + b\sum X \quad \dots \text{(Normal Equation I)}}$$ $$\frac{\partial S}{\partial b} = -2 \sum X(Y - a - bX) = 0 \implies \mathbf{\sum XY = a\sum X + b\sum X^2 \quad \dots \text{(Normal Equation II)}}$$

Shortcut Method (Origin Shifting):

By selecting the origin at the midpoint of the time series such that $\sum X = 0$, the normal equations simplify immediately:

$$\sum Y = Na + b(0) \implies \mathbf{a = \frac{\sum Y}{N}}$$ $$\sum XY = a(0) + b\sum X^2 \implies \mathbf{b = \frac{\sum XY}{\sum X^2}}$$
2. Complete Step-by-Step Numerical Computation
Exam Problem: The following table gives the annual production of a steel manufacturing plant in thousands of metric tonnes from 2016 to 2022 ($N = 7$ years). Fit a straight line trend equation $Y = a + bX$ by the method of least squares and estimate the production for the years 2024 and 2025.

Since $N = 7$ (odd number of years), we choose the middle year 2019 as the origin ($X = \text{Year} - 2019$ with unit $= 1$ year). This guarantees $\sum X = 0$.

Year ($t$) Production $Y$ ('000 tonnes) Coded Time $X = t - 2019$ $X^2$ $XY$ Trend Value $\hat{Y} = 88 + 4.61X$
2016 74 $-3$ 9 $-222$ $88 + 4.6071(-3) = 74.18$
2017 80 $-2$ 4 $-160$ $88 + 4.6071(-2) = 78.79$
2018 82 $-1$ 1 $-82$ $88 + 4.6071(-1) = 83.39$
2019 (Origin) 89 $0$ 0 $0$ $88 + 4.6071(0) = 88.00$
2020 91 $+1$ 1 $+91$ $88 + 4.6071(1) = 92.61$
2021 98 $+2$ 4 $+196$ $88 + 4.6071(2) = 97.21$
2022 102 $+3$ 9 $+306$ $88 + 4.6071(3) = 101.82$
Total ($N=7$) $\sum Y = 616$ $\sum X = 0$ $\sum X^2 = 28$ $\sum XY = 129$ $\sum \hat{Y} = 616.0$

Step 1: Compute Intercept $a$ and Slope $b$:

$$a = \frac{\sum Y}{N} = \frac{616}{7} = \mathbf{88.0}$$ $$b = \frac{\sum XY}{\sum X^2} = \frac{129}{28} \approx \mathbf{4.6071 \approx 4.61}$$

The fitted straight-line trend equation is:

$$\mathbf{\hat{Y} = 88.0 + 4.61 X \quad \text{(with origin: 2019, } X \text{ unit: 1 year)}}$$

Step 2: Forecast Future Production:

  • For Year 2024: $X = 2024 - 2019 = +5$ $$\hat{Y}_{2024} = 88.0 + 4.6071(5) = 88.0 + 23.04 = \mathbf{111.04 \text{ thousand metric tonnes}}$$
  • For Year 2025: $X = 2025 - 2019 = +6$ $$\hat{Y}_{2025} = 88.0 + 4.6071(6) = 88.0 + 27.64 = \mathbf{115.64 \text{ thousand metric tonnes}}$$

Q9. Explain the theory and construction of Shewhart $\bar{X}$ and $R$ Control Charts. Calculate control limits for 10 subgroups of size $n=5$, plot the $\bar{X}$ chart, and determine if the process is in statistical control.

🔥 Repeated: 2017, 2019, 2021, 2022 15 Marks
✍️ Comprehensive Solution:
1. Theoretical Framework of Control Charts

Developed by Dr. Walter A. Shewhart at Bell Telephone Laboratories, control charts distinguish between Chance (Random) Causes and Assignable (Special) Causes of variation:

  • $\bar{X}$ Chart (Chart for Means): Controls the process central tendency (process average or centering). Monitors whether machine tool wear, setting slippage, or temperature drift has shifted the process mean.
  • $R$ Chart (Chart for Ranges): Controls process dispersion or variability. Monitors consistency, operator uniformity, and machine vibration.

3-Sigma Control Limit Formulas:

Control Chart Lower Control Limit ($LCL$) Central Line ($CL$) Upper Control Limit ($UCL$)
$\bar{X}$ Chart $LCL_{\bar{X}} = \bar{\bar{X}} - A_2 \bar{R}$ $CL_{\bar{X}} = \bar{\bar{X}} = \frac{\sum \bar{X}}{k}$ $UCL_{\bar{X}} = \bar{\bar{X}} + A_2 \bar{R}$
$R$ Chart $LCL_R = D_3 \bar{R}$ $CL_R = \bar{R} = \frac{\sum R}{k}$ $UCL_R = D_4 \bar{R}$

For sample size $n = 5$: $A_2 = 0.577$, $D_3 = 0$, $D_4 = 2.114$.

2. Complete Step-by-Step Numerical Computation
Exam Problem: The following table shows the sample mean ($\bar{X}$) and range ($R$) of 10 subgroups of machined steel shafts, each subgroup containing $n = 5$ items:
Subgroup No. ($i$) 1 2 3 4 5 6 7 8 9 10 Total ($\sum$)
Sample Mean $\bar{X}_i$ (mm) 49.2 51.5 48.0 52.8 50.4 47.1 53.2 56.4 49.8 50.5 $\sum \bar{X} = \mathbf{508.9}$
Sample Range $R_i$ (mm) 6.2 7.0 5.8 8.1 6.5 7.4 6.9 9.2 6.0 6.9 $\sum R = \mathbf{70.0}$

Step 1: Compute Grand Mean ($\bar{\bar{X}}$) and Mean Range ($\bar{R}$):

$$\bar{\bar{X}} = \frac{\sum \bar{X}}{k} = \frac{508.9}{10} = \mathbf{50.89 \text{ mm}}$$ $$\bar{R} = \frac{\sum R}{k} = \frac{70.0}{10} = \mathbf{7.00 \text{ mm}}$$

Step 2: Calculate Control Limits for $\bar{X}$ Chart ($n=5, A_2 = 0.577$):

$$CL_{\bar{X}} = \bar{\bar{X}} = \mathbf{50.89}$$ $$UCL_{\bar{X}} = \bar{\bar{X}} + A_2 \bar{R} = 50.89 + (0.577 \times 7.00) = 50.89 + 4.039 = \mathbf{54.93 \text{ mm}}$$ $$LCL_{\bar{X}} = \bar{\bar{X}} - A_2 \bar{R} = 50.89 - (0.577 \times 7.00) = 50.89 - 4.039 = \mathbf{46.85 \text{ mm}}$$

Step 3: Calculate Control Limits for $R$ Chart ($n=5, D_3 = 0, D_4 = 2.114$):

$$CL_R = \bar{R} = \mathbf{7.00}$$ $$UCL_R = D_4 \bar{R} = 2.114 \times 7.00 = \mathbf{14.80 \text{ mm}}$$ $$LCL_R = D_3 \bar{R} = 0 \times 7.00 = \mathbf{0.00 \text{ mm}}$$
3. Vector SVG Diagram: Shewhart $\bar{X}$ Control Chart
FIGURE 5: SHEWHART $\bar{X}$ CONTROL CHART (3-SIGMA LIMITS & ASSIGNABLE CAUSE OUTLIER)
PROCESS CONTROL MONITORING: SAMPLE MEANS ($\bar{X}$ CHART) Sample Size $n=5$ • Grand Mean $\bar{\bar{X}} = 50.89$ • 3-Sigma Limits $[46.85, 54.93]$ 54.93 50.89 46.85 Sample Mean $\bar{X}$ (mm) UCL = 54.93 mm CL = 50.89 mm LCL = 46.85 mm 1 2 3 4 5 6 7 8* Sample 8 (56.4 > UCL 54.93) • Out of Control! 9 10 Subgroup Number ($k = 10$ Samples)

Step 4: Statistical Decision and Process Diagnosis:

Process Conclusion:
  • On the $\bar{X}$ chart, Subgroup 8 ($\bar{X}_8 = 56.4 \text{ mm}$) exceeds the Upper Control Limit ($UCL = 54.93 \text{ mm}$).
  • On the $R$ chart, all ranges lie well within limits ($0 \le R_i \le 14.80$).
  • Final Verdict: The process is NOT in statistical control. An assignable cause of variation shifted the process average upwards during sample 8 (such as tool slippage, human setting error, or sudden power spike). Management must investigate and eliminate this assignable cause before recalculating revised control limits.

Q10. Explain the concepts of Skewness and Kurtosis. Differentiate between Symmetrical, Positively Skewed, and Negatively Skewed frequency distributions with diagrams and formulas (Karl Pearson, Bowley, and Central Moments).

🔥 Repeated: 2017, 2018, 2020, 2022 15 Marks
✍️ Comprehensive Solution:
1. Nature and Measurement of Skewness

Skewness refers to the lack of symmetry (asymmetry) in a frequency distribution curve. While central tendency measures the location of data and dispersion measures its spread, skewness describes the direction and degree of departure from symmetry.

  • Symmetrical Distribution ($S_k = 0$): The curve is balanced bell-shaped. Mean, Median, and Mode coincide at the same point: $$\mathbf{\text{Mean} = \text{Median} = \text{Mode}}$$ Frequencies are evenly distributed on both sides of the central peak. Quartiles are equidistant from median: $Q_3 - M = M - Q_1$.
  • Positively Skewed Distribution ($S_k > 0$, Right-Tailed): The tail of the curve stretches out further to the right (positive direction). The longer right tail pulls the Mean towards larger values: $$\mathbf{\text{Mean} > \text{Median} > \text{Mode}}$$ Quartile property: $(Q_3 - M) > (M - Q_1)$. Common in income distributions, wealth holdings, and city house prices.
  • Negatively Skewed Distribution ($S_k < 0$, Left-Tailed): The tail of the curve stretches out further to the left (negative direction). Low extreme values pull the Mean to the left: $$\mathbf{\text{Mean} < \text{Median} < \text{Mode}}$$ Quartile property: $(Q_3 - M) < (M - Q_1)$. Common in human age at death or exam scores on an easy paper.
2. Vector SVG Diagram: Skewness Taxonomy
FIGURE 6: FREQUENCY CURVES AND MEASURES OF SKEWNESS
TAXONOMY OF ASYMMETRY: FREQUENCY CURVES & CENTRAL TENDENCY ALIGNMENT SYMMETRICAL ($S_k = 0$) Normal / Bell-Shaped Curve $\bar{X} = M = Z$ Mean = Median = Mode POSITIVELY SKEWED ($S_k > 0$) Right-Tailed Distribution Mode Med Mean $\text{Mode} < \text{Median} < \text{Mean}$ Long tail on positive side NEGATIVELY SKEWED ($S_k < 0$) Left-Tailed Distribution Mean Med Mode $\text{Mean} < \text{Median} < \text{Mode}$ Long tail on negative side
3. Standard Measures of Relative Skewness
Method Formula Range & Interpretation
Karl Pearson's Coefficient ($S_k$) $$S_k = \frac{\text{Mean} - \text{Mode}}{\sigma} = \frac{3(\text{Mean} - \text{Median})}{\sigma}$$ Ranges theoretically between $-3$ and $+3$. In unimodal practical distributions, lies between $-1$ and $+1$.
Bowley's Quartile Coefficient ($S_q$) $$S_q = \frac{(Q_3 - M) - (M - Q_1)}{(Q_3 - M) + (M - Q_1)} = \frac{Q_3 + Q_1 - 2M}{Q_3 - Q_1}$$ Strictly bounded between $-1$ and $+1$. Ideal for open-end class intervals because it relies solely on quartiles.
Moment Coefficient ($\beta_1, \gamma_1$) $$\beta_1 = \frac{\mu_3^2}{\mu_2^3}, \quad \gamma_1 = \sqrt{\beta_1} = \frac{\mu_3}{\sigma^3}$$ For symmetrical curve: $\mu_3 = 0 \implies \gamma_1 = 0$. If $\gamma_1 > 0$, positive skew; if $\gamma_1 < 0$, negative skew.
4. Concept and Measures of Kurtosis

Kurtosis (from Greek kyrtos meaning convex) measures the peakedness or flatness of a unimodal frequency distribution relative to a standard Normal Distribution:

  • 1. Mesokurtic Curve ($\beta_2 = 3, \gamma_2 = 0$): The standard normal bell curve with normal peakedness.
  • 2. Leptokurtic Curve ($\beta_2 > 3, \gamma_2 > 0$): More slender, sharply peaked with heavier tails than normal curve.
  • 3. Platykurtic Curve ($\beta_2 < 3, \gamma_2 < 0$): Flatter-topped, broader spread with lighter tails than normal curve.

Formula for Kurtosis based on central moments:

$$\mathbf{\beta_2 = \frac{\mu_4}{\mu_2^2} = \frac{\mu_4}{\sigma^4}, \quad \gamma_2 = \beta_2 - 3}$$

Q11. Define the Poisson Distribution. State its conditions and PMF. Prove that its Mean and Variance are both equal to $\lambda$, and solve a complete industrial manufacturing defect problem.

🔥 Repeated: 2017, 2018, 2019, 2021, 2022 15 Marks
✍️ Comprehensive Solution:
1. Definition and Limiting Conditions

The Poisson Distribution, discovered by French mathematician Siméon Denis Poisson in 1837, is a discrete probability distribution representing the count of rare events occurring randomly in a fixed interval of time or space.

Limiting Form of the Binomial Distribution:

The Binomial distribution $B(n, p)$ tends to a Poisson distribution under the following three limiting conditions:

  1. The number of trials $n$ is indefinitely large ($n \to \infty$).
  2. The constant probability of success $p$ is extremely small ($p \to 0$).
  3. The product $np = \lambda$ remains finite and constant ($\lambda > 0$).

Probability Mass Function (PMF):

$$\mathbf{P(X = x) = \frac{e^{-\lambda} \lambda^x}{x!}, \quad x = 0, 1, 2, 3, \dots, \infty}$$

where $e \approx 2.71828$ is Euler's constant and $\lambda$ is the mean parameter.

2. Mathematical Proof: Sum of Probabilities = 1
$$\sum_{x=0}^{\infty} P(X = x) = \sum_{x=0}^{\infty} \frac{e^{-\lambda} \lambda^x}{x!} = e^{-\lambda} \sum_{x=0}^{\infty} \frac{\lambda^x}{x!} = e^{-\lambda} \left(1 + \lambda + \frac{\lambda^2}{2!} + \frac{\lambda^3}{3!} + \dots\right)$$

By Maclaurin expansion of $e^{\lambda}$:

$$\sum_{x=0}^{\infty} P(X = x) = e^{-\lambda} \cdot e^{\lambda} = e^0 = \mathbf{1} \quad \blacksquare$$
3. Mathematical Derivations of Mean and Variance

Proof for Mean $E(X)$:

$$E(X) = \sum_{x=0}^{\infty} x \cdot P(X = x) = \sum_{x=0}^{\infty} x \cdot \frac{e^{-\lambda} \lambda^x}{x!} = \sum_{x=1}^{\infty} \frac{e^{-\lambda} \lambda^x}{(x-1)!}$$

Factoring out $\lambda e^{-\lambda}$:

$$E(X) = \lambda e^{-\lambda} \sum_{x=1}^{\infty} \frac{\lambda^{x-1}}{(x-1)!} = \lambda e^{-\lambda} \left(1 + \lambda + \frac{\lambda^2}{2!} + \dots\right) = \lambda e^{-\lambda} \cdot e^{\lambda} = \mathbf{\lambda}$$

Hence, Mean of Poisson distribution = $\lambda$. $\blacksquare$

Proof for Variance $Var(X)$:

First find $E(X(X-1))$:

$$E(X(X-1)) = \sum_{x=0}^{\infty} x(x-1) \frac{e^{-\lambda} \lambda^x}{x!} = \sum_{x=2}^{\infty} \frac{e^{-\lambda} \lambda^x}{(x-2)!} = \lambda^2 e^{-\lambda} \sum_{x=2}^{\infty} \frac{\lambda^{x-2}}{(x-2)!} = \lambda^2 e^{-\lambda} \cdot e^{\lambda} = \lambda^2$$

Now express $E(X^2)$ as $E(X(X-1)) + E(X)$:

$$E(X^2) = \lambda^2 + \lambda$$

Therefore, Variance is:

$$Var(X) = E(X^2) - [E(X)]^2 = (\lambda^2 + \lambda) - \lambda^2 = \mathbf{\lambda}$$
Distinctive Axiom: The Poisson distribution is the only discrete probability distribution where $\text{Mean} = \text{Variance} = \lambda$.
4. Complete Step-by-Step Numerical Solution
Exam Problem: A manufacturer of precision computer chips finds that 2% of the chips produced are defective. In a random sample box of 100 chips, find the probability that:
  1. No chip is defective ($x = 0$)
  2. Exactly 2 chips are defective ($x = 2$)
  3. At most 2 chips are defective ($x \le 2$)
  4. At least 1 chip is defective ($x \ge 1$)
(Given: $e^{-2} \approx 0.1353$)

Step 1: Identify Parameters:

  • Sample size $n = 100$ (large)
  • Defective probability $p = 0.02$ (small)
  • Poisson Parameter $\lambda = np = 100 \times 0.02 = \mathbf{2.0}$

Step 2: Probability of No Defective Chips ($x = 0$):

$$P(X = 0) = \frac{e^{-2} \cdot 2^0}{0!} = \frac{0.1353 \times 1}{1} = \mathbf{0.1353 \text{ (or } 13.53\%)}$$

Step 3: Probability of Exactly 2 Defective Chips ($x = 2$):

$$P(X = 2) = \frac{e^{-2} \cdot 2^2}{2!} = \frac{0.1353 \times 4}{2} = 0.1353 \times 2 = \mathbf{0.2706 \text{ (or } 27.06\%)}$$

Step 4: Probability of At Most 2 Defective Chips ($X \le 2$):

$$P(X \le 2) = P(X = 0) + P(X = 1) + P(X = 2)$$ $$P(X = 1) = \frac{e^{-2} \cdot 2^1}{1!} = 0.1353 \times 2 = 0.2706$$ $$P(X \le 2) = 0.1353 + 0.2706 + 0.2706 = \mathbf{0.6765 \text{ (or } 67.65\%)}$$

Step 5: Probability of At Least 1 Defective Chip ($X \ge 1$):

$$P(X \ge 1) = 1 - P(X = 0) = 1 - 0.1353 = \mathbf{0.8647 \text{ (or } 86.47\%)}$$

Q12. Explain the Multiplicative and Additive Models of Time Series Decomposition. Detail the four components, and calculate 4-Year Centered Moving Averages for a 10-year dataset.

🔥 Repeated: 2017, 2018, 2020, 2021, 2022 15 Marks
✍️ Comprehensive Solution:
1. Mathematical Models of Time Series Decomposition

A time series $Y_t$ is a sequence of observations recorded over successive, equally spaced time intervals. Classical time series analysis assumes that $Y_t$ is the resultant of four distinct underlying forces:

  • Secular Trend ($T_t$): Long-term smooth growth or decline over decades.
  • Seasonal Variation ($S_t$): Intra-year periodic movements repeating every 12 months or 4 quarters.
  • Cyclical Fluctuations ($C_t$): Multi-year oscillatory macroeconomic business cycles (wave-like).
  • Irregular / Random Shocks ($I_t$): Unpredictable, erratic residual disturbances (floods, wars, strikes).
Model Mathematical Equation Underlying Assumption & Usage
Additive Model $$\mathbf{Y_t = T_t + S_t + C_t + I_t}$$ Assumes the four components act independently in absolute physical units. The seasonal variation magnitude does not depend on trend level.
Multiplicative Model (Standard) $$\mathbf{Y_t = T_t \times S_t \times C_t \times I_t}$$ Assumes the components are interdependent. $T_t$ is in original units while $S_t, C_t, I_t$ are dimensionless ratios or percentages. Widely used in business & economics.
2. Vector SVG Diagram: Time Series 4-Component Decomposition
FIGURE 7: CLASSICAL 4-COMPONENT DECOMPOSITION ARCHITECTURE OF TIME SERIES
TIME SERIES DECOMPOSITION: COMPOSITE OBSERVED DATA ($Y_t = T \times S \times C \times I$) Observed Series ($Y_t$): Raw Data = Trend + Seasons + Cycles + Noise T 1. Secular Trend ($T_t$) Smooth, persistent long-term growth/decay Steady Uptrend ↗ S 2. Seasonal Variation ($S_t$) Periodic intra-year pattern (< 12 months) Quarterly Peaks • Period $\le 1$ Yr 3. Cyclical Fluctuations ($C_t$) Wave-like business cycles (3 to 10 years) Boom → Recession → Recovery I 4. Irregular / Random Noise ($I_t$) Unforeseen shocks: floods, strikes, wars Unpredictable Residual Jitter Multiplicative: $Y = T \times S \times C \times I$  |  Additive: $Y = T + S + C + I$
3. Moving Average Method & 4-Year Centering Algorithm

The Moving Average method smooths out short-term seasonal and irregular fluctuations to reveal the underlying secular trend. When the moving average period is even (e.g., $m = 4$ years), the 4-year average falls between two calendar years (e.g., midway between year 2 and 3). To align the trend values exactly with actual calendar years, a second 2-period centering moving average is computed.

4. Complete Step-by-Step Numerical Computation
Exam Problem: Calculate 4-Year Centered Moving Averages for the following annual net profit data (in lakhs of rupees) from 2013 to 2022:
Year Net Profit $Y$ (₹ Lakhs) 4-Year Moving Total (placed midway) 4-Year Centered Total (Sum of two 4-yr totals) 4-Year Centered Moving Average (Trend = Total / 8)
2013 36 — — —
2014 42 $36+42+40+46 = \mathbf{164}$ — —
2015 40 $164 + 176 = \mathbf{340}$ $340 / 8 = \mathbf{42.50}$
$42+40+46+48 = \mathbf{176}$
2016 46 $176 + 186 = \mathbf{362}$ $362 / 8 = \mathbf{45.25}$
$40+46+48+52 = \mathbf{186}$
2017 48 $186 + 197 = \mathbf{383}$ $383 / 8 = \mathbf{47.88}$
$46+48+52+51 = \mathbf{197}$
2018 52 $197 + 207 = \mathbf{404}$ $404 / 8 = \mathbf{50.50}$
$48+52+51+56 = \mathbf{207}$
2019 51 $207 + 219 = \mathbf{426}$ $426 / 8 = \mathbf{53.25}$
$52+51+56+60 = \mathbf{219}$
2020 56 $219 + 233 = \mathbf{452}$ $452 / 8 = \mathbf{56.50}$
$51+56+60+66 = \mathbf{233}$
2021 60 — — —
2022 66 — — —

Methodological Insight:

In 4-year centering, the first two years (2013, 2014) and the last two years (2021, 2022) do not have centered trend values because a full 4-year window cannot be centered around them. The centered trend values show a smooth, unambiguous upward progression from ₹42.50 Lakhs in 2015 to ₹56.50 Lakhs in 2020.

HIGH-FREQUENCY REVISION

Comprehensive Statistical Formula Cheat Sheet

Instant Reference of Essential Formulas for CCSU BCA-305 University Examination

Statistical Topic Standard Mathematical Formula Key Mathematical Properties & Exam Tips
Arithmetic Mean ($\bar{X}$) $$\bar{X} = A + \frac{\sum f d}{N} \times h \quad \left(d = \frac{X - A}{h}\right)$$ $\sum (X - \bar{X}) = 0$; $\sum (X - \bar{X})^2$ is minimum. Affected by change of origin and scale.
Median ($M$) & Mode ($Z$) $$M = L_1 + \frac{\frac{N}{2} - C}{f} \times h, \quad Z = L_1 + \frac{f_1 - f_0}{2f_1 - f_0 - f_2} \times h$$ Empirical relation: $\text{Mode} \approx 3\,\text{Median} - 2\,\text{Mean}$. Median obtained graphically via Ogives intersection.
Standard Deviation ($\sigma$) & C.V. $$\sigma = \sqrt{\frac{\sum f d^2}{N} - \left(\frac{\sum f d}{N}\right)^2} \times h, \quad C.V. = \frac{\sigma}{\bar{X}} \times 100$$ Independent of change of origin; dependent on change of scale. Lower C.V. implies higher consistency/stability.
Karl Pearson's Correlation ($r$) $$r = \frac{N\sum XY - (\sum X)(\sum Y)}{\sqrt{[N\sum X^2 - (\sum X)^2][N\sum Y^2 - (\sum Y)^2]}}$$ Strict limits: $-1 \le r \le +1$. Pure dimensionless number. Completely independent of both origin and scale changes.
Spearman's Rank Correlation ($\rho$) $$\rho = 1 - \frac{6 \sum D^2}{N(N^2 - 1)}$$ Used for qualitative ranks. For tied ranks: add $\frac{m(m^2-1)}{12}$ correction term for each tied group.
Regression Coefficients & Lines $$b_{yx} = r \frac{\sigma_y}{\sigma_x}, \quad b_{xy} = r \frac{\sigma_x}{\sigma_y}, \quad r = \pm \sqrt{b_{yx} \times b_{xy}}$$ Both coefficients have the same sign as $r$. If one is $>1$, the other must be $<1$. The arithmetic mean of $b_{yx}$ and $b_{xy}$ is $\ge r$. Lines cross at $(\bar{X}, \bar{Y})$.
Fisher's Ideal Index ($P_{01}^F$) $$P_{01}^F = \sqrt{\frac{\sum p_1 q_0}{\sum p_0 q_0} \times \frac{\sum p_1 q_1}{\sum p_0 q_1}} \times 100$$ Geometric mean of Laspeyres and Paasche. Satisfies Time Reversal Test ($P_{01} \times P_{10} = 1$) and Factor Reversal Test ($P_{01} \times Q_{01} = V_{01}$).
Poisson Distribution PMF $$P(X = x) = \frac{e^{-\lambda} \lambda^x}{x!}, \quad x = 0, 1, 2, \dots$$ Limiting form of Binomial ($n \to \infty, p \to 0, np = \lambda$). Unique property: $\text{Mean} = \text{Variance} = \lambda$.
Normal Distribution Standard Variate $$Z = \frac{X - \mu}{\sigma} \sim N(0, 1), \quad f(z) = \frac{1}{\sqrt{2\pi}} e^{-z^2/2}$$ Symmetrical bell curve. Empirical Rule: $68.27\%$ within $\mu \pm \sigma$, $95.45\%$ within $\mu \pm 2\sigma$, $99.73\%$ within $\mu \pm 3\sigma$.
$\bar{X}$ and $R$ Control Limits (SQC) $$UCL_{\bar{X}} = \bar{\bar{X}} + A_2 \bar{R}, \quad LCL_{\bar{X}} = \bar{\bar{X}} - A_2 \bar{R}$$ Controls process mean and variability. Distinguishes between Chance variation (stable) and Assignable causes (out of control).
TOPPERS' BLUEPRINT

Strategic Guidelines for Scoring 70+/75 in Elements of Statistics

Proven Techniques for Maximum Marks in CCSU University Evaluations

1. Write Formula First & Define Symbols

Always state the complete theoretical formula and define every symbol (e.g., $N, \bar{X}, \sigma, h$) before substituting numbers. University evaluators award 2 to 3 marks purely for the correct formula even if minor arithmetic slips happen later.

2. Structured Tabular Columns with Headers

In numerical questions (Karl Pearson correlation, standard deviation, time series least squares), draw crisp tables with precise column headers ($X, Y, d_x, d_y, d_x^2, d_y^2, d_x d_y$). Double-check column totals before formula substitution.

3. Sketch Clean Statistical Visuals

For Section C long answers, always sketch neat diagrams: the Bell curve with percentage areas for Normal distribution, intersecting Ogives for Median, and control limits ($UCL, CL, LCL$) for SQC. Visual answers command top tier marks.

4. State Concrete Business Interpretations

Never stop at a raw numerical value. Conclude with an interpretation statement: e.g., "Since $r = +0.89$, there exists a strong positive correlation", or "Sample 8 exceeds UCL, hence process is out of statistical control due to assignable causes."