Theory of Attributes

Unit V · Lecture 40

BDS113 — Introduction to Statistics

Unit V Review & Problem-Solving Session

Formula Sheet · Concept Map · 8 Worked Exam-Style Problems · Common Mistakes · Comprehensive Quiz · Course Wrap-Up

Unit V · Lecture 40 of 40 · B.Sc. Data Science, 1st Year, Semester II · Final Lecture!

Theory of Attributes

Unit V · Lecture 40

The Final Review Journey

📋

Complete Formula Sheet — Unit V
Every formula from Lectures 34–39 in one place

🗺️

Concept Map & Decision Guide
How all Unit V topics connect · When to use which measure

✏️

8 Worked Exam-Style Problems (Click to Reveal)
Covering every topic · Step-by-step solutions · Exam technique

🎓

Common Mistakes, Exam Tips, Quiz & Course Wrap-Up
Pitfalls to avoid · 12-question quiz · Complete course summary

Theory of Attributes

Unit V · Lecture 40

📋 Complete Formula Sheet — Unit V

Notation & Identities (Lecture 34)

(A) = (AB) + (Aβ) | (B) = (AB) + (αB) | N = (A) + (α) = (B) + (β)

Consistency Conditions (Lecture 35)

max(0, (A)+(B)−N) ≤ (AB) ≤ min((A), (B)) | All cross-class frequencies ≥ 0

Independence Test (Lecture 36)

Independent: (AB) = (A)(B)/N | Positive: (AB) > (A)(B)/N | Negative: (AB) < (A)(B)/N

Chi-Square & Coefficients (Lecture 37)

χ² = N(ad−bc)² / [(a+b)(c+d)(a+c)(b+d)] | φ = √(χ²/N) | C = √(χ²/(χ²+N))

Yule's Q (Lecture 38)

Q = (ad − bc) / (ad + bc) | Range: −1 to +1 | Q = 0 iff ad = bc (independence)

Yule's Y (Lecture 39)

Y = (√ad − √bc) / (√ad + √bc) | |Y| ≤ |Q| always | Q = 2Y/(1+Y²)

2×2 Table (Standard Layout)

       B     β    Total

    A     a     b    (A)

    α     c     d    (α)

    Total (B)   (β)   N

Theory of Attributes

Unit V · Lecture 40

🗺️ Concept Map & Decision Guide

1 Start: You have qualitative (attribute) data in a 2×2 table.
Identify the two attributes A and B. Label the cells a, b, c, d. Compute row and column totals.

2 Check Consistency (Lecture 35):
Verify max(0, (A)+(B)−N) ≤ (AB) ≤ min((A), (B)). If violated → data is impossible.

3 Test Independence (Lecture 36):
Compute E(AB) = (A)(B)/N. Compare with observed (AB). Equal → independent. Different → associated.

4 Measure Strength of Association:
→ 2×2 table: Use Yule's Q (most common in exams) or Yule's Y
→ Any r×c table: Use Pearson's C
→ For hypothesis testing: Use χ² and φ

5 Interpret:
State direction (positive/negative), strength (weak/moderate/strong), and practical meaning in context.

💬 Exam Shortcut: Most exam problems on Unit V follow this exact sequence: (1) build the 2×2 table, (2) check consistency, (3) test independence, (4) compute Q or Y, (5) interpret. Master this flow and you can solve any Unit V problem.

Theory of Attributes

Unit V · Lecture 40

✏️ Problem 1: Build Table & Check Consistency

📝 In a survey of 400 people: 250 are vaccinated (V), 150 are not. 200 are above 50 (O), 200 are not. 130 are vaccinated AND above 50. (a) Construct the 2×2 table. (b) Is the data consistent?

▶ Reveal

(a) Build the table:

(V) = 250, (O) = 200, (VO) = 130, N = 400
(Vω) = 250 − 130 = 120
(νO) = 200 − 130 = 70
(νω) = 400 − 250 − 200 + 130 = 80 [= 150 − 70 = 80 ✓]

	O	ω	Total
V	130	120	250
ν	70	80	150
Total	200	200	400

(b) Consistency check:

Lower bound: max(0, 250+200−400) = max(0, 50) = 50
Upper bound: min(250, 200) = 200
(VO) = 130. Is 50 ≤ 130 ≤ 200? Yes ✓
All cross-class frequencies (130, 120, 70, 80) ≥ 0. Data is CONSISTENT ✓

Theory of Attributes

Unit V · Lecture 40

✏️ Problem 2: Inconsistent Data

📝 Given N = 300, (A) = 200, (B) = 180, (AB) = 30. Check consistency. If inconsistent, find the minimum (AB) must be.

▶ Reveal

Lower bound: max(0, 200+180−300) = max(0, 80) = 80
Upper bound: min(200, 180) = 180

(AB) = 30. Is 80 ≤ 30? NO ✗

Data is INCONSISTENT.

Verification: (αβ) = 300 − 200 − 180 + 30 = −50 (negative — impossible!)

Minimum (AB) must be 80 for consistency.
At (AB) = 80: (αβ) = 300 − 200 − 180 + 80 = 0 (just barely consistent).

Theory of Attributes

Unit V · Lecture 40

✏️ Problem 3: Test for Independence

📝 Given: a=80, b=40, c=60, d=120, N=300. (a) Test independence. (b) If not independent, state the type of association.

▶ Reveal

(a) Expected frequency under independence:

(A) = 80+40 = 120, (B) = 80+60 = 140
E(AB) = (A)(B)/N = 120×140/300 = 56

Observed (AB) = 80

Since 80 ≠ 56, the attributes are NOT independent.

(b) Type of association:

(AB) = 80 > 56 = E(AB) → Positive association

The attributes A and B occur together more often than expected by chance.

Theory of Attributes

Unit V · Lecture 40

✏️ Problem 4: Yule's Q — Complete Analysis

📝 A study of 500 patients recorded treatment (A) and recovery (B): (AB)=150, (Aβ)=100, (αB)=50, (αβ)=200. Compute Q and interpret.

▶ Reveal

a=150, b=100, c=50, d=200

Q = (ad − bc) / (ad + bc)
   = (150×200 − 100×50) / (150×200 + 100×50)
   = (30000 − 5000) / (30000 + 5000)
   = 25000 / 35000
   = 0.7143

Interpretation: Q = 0.714 indicates a strong positive association between treatment and recovery. Patients who received treatment were much more likely to recover than those who did not.

Verification — Independence check:

E(AB) = 250×200/500 = 100. Observed = 150 > 100. Confirms positive association ✓

Theory of Attributes

Unit V · Lecture 40

✏️ Problem 5: Yule's Y & Comparison with Q

📝 For the same data (a=150, b=100, c=50, d=200), compute Yule's Y and compare with Q.

▶ Reveal

√ad = √(150×200) = √30000 = 173.21
√bc = √(100×50) = √5000 = 70.71

Y = (√ad − √bc) / (√ad + √bc)
   = (173.21 − 70.71) / (173.21 + 70.71)
   = 102.50 / 243.92
   = 0.4202

Measure	Value	Strength
Q	0.714	Strong positive
Y	0.420	Moderate positive

|Y| = 0.420 ≤ |Q| = 0.714 ✓ (The inequality always holds.)

Q and Y agree on direction (positive) but differ on strength. Q is more sensitive to the large diagonal products.

Theory of Attributes

Unit V · Lecture 40

✏️ Problem 6: χ², φ, and C

📝 For a=150, b=100, c=50, d=200, N=500, compute χ², φ, and C.

▶ Reveal

χ² = N(ad−bc)² / [(a+b)(c+d)(a+c)(b+d)]
   = 500(30000−5000)² / [250×250×200×300]
   = 500(25000)² / [250×250×200×300]
   = 500 × 625,000,000 / 3,750,000,000
   = 312,500,000,000 / 3,750,000,000
   = 83.33

φ = √(χ²/N) = √(83.33/500) = √0.1667 = 0.4082

C = √(χ²/(χ²+N)) = √(83.33/583.33) = √0.1429 = 0.3780

Measure	Value	Interpretation
χ²	83.33	Large discrepancy from independence
φ	0.408	Moderate association
C	0.378	Moderate association (C < φ ✓)
Q	0.714	Strong positive association
Y	0.420	Moderate positive association

Key observation: Different measures give different "strength" labels for the same data. This is normal — each measure captures association differently. Always state which measure you're using.

Theory of Attributes

Unit V · Lecture 40

✏️ Problem 7: Finding a Missing Frequency

📝 Given N=400, (A)=200, (B)=180, and Q = 0.5. Find (AB).

▶ Reveal

Let (AB) = x. Then:
a = x, b = 200−x, c = 180−x, d = 400−200−180+x = 20+x

Q = (ad − bc) / (ad + bc) = 0.5

ad = x(20+x) = 20x + x²
bc = (200−x)(180−x) = 36000 − 380x + x²

ad − bc = 20x + x² − 36000 + 380x − x² = 400x − 36000
ad + bc = 20x + x² + 36000 − 380x + x² = 2x² − 360x + 36000

(400x − 36000) / (2x² − 360x + 36000) = 0.5

400x − 36000 = 0.5(2x² − 360x + 36000)
400x − 36000 = x² − 180x + 18000
x² − 580x + 54000 = 0

Using quadratic formula:
x = [580 ± √(336400 − 216000)] / 2
= [580 ± √120400] / 2
= [580 ± 347.0] / 2

x = (580+347)/2 = 463.5 → too large (exceeds min(A,B)=180)
x = (580−347)/2 = 116.5 → valid (within consistency range)

(AB) ≈ 117 (rounding to nearest integer)

Verification: Consistency range: max(0, 200+180−400)=0 to min(200,180)=180. 117 is in range ✓

Theory of Attributes

Unit V · Lecture 40

✏️ Problem 8: Three Attributes — Consistency & Independence

📝 Given N=1000, (A)=400, (B)=500, (C)=300, (AB)=200, (AC)=150, (BC)=180, (ABC)=80. (a) Check pairwise consistency. (b) Check mutual independence.

▶ Reveal

(a) Pairwise consistency:

A–B: max(0, 400+500−1000)=0 to min(400,500)=400. (AB)=200 ∈ [0,400] ✓
A–C: max(0, 400+300−1000)=0 to min(400,300)=300. (AC)=150 ∈ [0,300] ✓
B–C: max(0, 500+300−1000)=0 to min(500,300)=300. (BC)=180 ∈ [0,300] ✓

All pairwise checks pass → pairwise consistent ✓

(b) Mutual independence test:

If mutually independent:
E(AB) = 400×500/1000 = 200. Observed = 200. A–B: independent ✓
E(AC) = 400×300/1000 = 120. Observed = 150. A–C: NOT independent ✗ (positive)
E(BC) = 500×300/1000 = 150. Observed = 180. B–C: NOT independent ✗ (positive)

E(ABC) = (A)(B)(C)/N² = 400×500×300/1000000 = 60
Observed (ABC) = 80 ≠ 60. Triple: NOT independent ✗

Conclusion: A and B are pairwise independent, but A–C and B–C are positively associated. The three attributes are NOT mutually independent.

Theory of Attributes

Unit V · Lecture 40

⚠️ Common Mistakes in Unit V

Notation Errors

1. Confusing (A) with A — (A) is a frequency (number), A is the attribute itself.

2. Writing (AB) when you mean (Aβ) — read the notation carefully!

3. Forgetting that (α) = N − (A), not (N) − (A) — there is no (N) class.

4. Mixing up rows and columns in the 2×2 table — always use the standard layout.

Formula Errors

1. In Q = (ad−bc)/(ad+bc), forgetting that a,b,c,d are the four cells, not the marginals.

2. In Y, computing √(ad) − √(bc) vs √(a)·√(d) − √(b)·√(c) — they are the same! (√a·√d = √(ad)).

3. Forgetting that |Y| ≤ |Q| — use this to verify your answers.

4. In χ², using the wrong formula for 2×2 tables — memorise the shortcut N(ad−bc)²/[...].

Conceptual Errors

1. Thinking Q = 0 means the attributes are mutually exclusive (it means independent).

2. Thinking Q = 1 means all individuals have both attributes (it means perfect association).

3. Confusing consistency with independence — they are different concepts!

4. Not interpreting the result — always state direction and strength.

💬 Exam Tip: After computing Q or Y, always do a quick sanity check: (1) Is |Q| ≤ 1? (2) Is |Y| ≤ |Q|? (3) Does the sign match the direction you'd expect from the data? (4) Are all cross-class frequencies non-negative? These catches prevent most errors.

Theory of Attributes

Unit V · Lecture 40

📝 Exam Tips for Unit V

📊

Always Draw the Table

Write the 2×2 table with a,b,c,d labels before computing anything. This prevents notation confusion.

✅

Verify Everything

Check: a+b+c+d=N. All cells ≥ 0. |Q|≤1. |Y|≤|Q|. These catches prevent most arithmetic errors.

🎯

Interpret Every Answer

Don't just give Q = 0.71. Say: "Strong positive association between A and B." Interpretation earns marks.

⏱️

Unit V is Quick

Unit V problems are shorter than Unit IV (no large computation tables). Use saved time for checking.

📖 High-Frequency Exam Questions — Unit V

1. Given partial information, construct the complete 2×2 table. (Very common)
2. Check consistency of given frequency data. (Common)
3. Test whether two attributes are independent. If not, state the type of association. (Very common)
4. Compute Yule's Q and interpret. (Almost every exam)
5. Compute Yule's Y and compare with Q. (Common)
6. Compute χ², φ, and/or C. (Occasional)
7. Given Q and partial data, find a missing frequency. (Occasional, harder)
8. State properties of Q or Y (range, special cases). (Theory question)

Theory of Attributes

Unit V · Lecture 40

🧠 Comprehensive Quiz — Unit V (12 Questions)

Q1: (A) represents:

A) The attribute A itself

B) The frequency (number) of individuals possessing A

C) The probability of A

D) The complement of A

Q2: (A) = (AB) + (Aβ) is:

A) Sometimes true

B) Always true (by definition of dichotomous classification)

C) True only when A and B are independent

D) True only for consistent data

Q3: N=200, (A)=120, (B)=140. Minimum (AB) for consistency:

A) 0

B) 20

C) 60

D) 120

Q4: If (AB) = (A)(B)/N, the attributes are:

A) Independent

B) Positively associated

C) Negatively associated

D) Mutually exclusive

Q5: For a=40, b=60, c=30, d=70, Q = ?

A) 0.25

B) 0.368

C) 0.50

D) 0.75

Q6: Q = +1 occurs when:

A) All individuals have both attributes

B) b = 0 or c = 0 (one off-diagonal cell is zero)

C) a = b = c = d

D) The attributes are independent

Q7: Yule's Y uses:

A) Products ad and bc

B) Geometric means √ad and √bc

C) Arithmetic means (a+d)/2 and (b+c)/2

D) Logarithms of ad and bc

Q8: The relationship between Q and Y is:

A) Q = Y always

B) |Y| ≤ |Q| always

C) |Q| ≤ |Y| always

D) Q + Y = 1

Q9: Pearson's C can be used for:

A) Only 2×2 tables

B) Any r×c contingency table

C) Only when attributes are independent

D) Only quantitative data

Q10: If (αβ) = −15, the data is:

A) Consistent with negative association

B) Inconsistent (negative frequency is impossible)

C) Independent

D) Perfectly associated

Q11: For the same 2×2 data, which is always true?

A) φ > C

B) C < φ (for non-zero χ²)

C) C = φ

D) C > Q

Q12: Three attributes A, B, C are mutually independent if:

A) Each pair is independent

B) All pairs are independent AND (ABC) = (A)(B)(C)/N²

C) (ABC) = 0

D) Q = 0 for all pairs

Theory of Attributes

Unit V · Lecture 40

BDS113 — Course Complete!

Introduction to Statistics · All 40 Lectures · 5 Units

Unit I

Descriptive Statistics
Measures of Central Tendency & Dispersion

Lectures 1–8

Unit II

Probability
Axioms, Conditional, Bayes' Theorem

Lectures 9–16

Unit III

Random Variables & Distributions
Binomial, Poisson, Normal

Lectures 17–23

Unit IV

Correlation & Regression
Pearson, Spearman, Regression Lines

Lectures 24–33

Unit V

Theory of Attributes
Independence, Q, Y, Contingency

Lectures 34–40

Key Skills You've Developed

Mean, Median, Mode Variance & SD Probability Bayes' Theorem Distributions Normal Curve Correlation Regression Curve Fitting Attributes Yule's Q & Y Contingency

You have completed all 40 lectures of BDS113.

Statistics is not just a subject — it is a way of thinking about the world with data. The tools you've learned here — from computing a mean to fitting a regression line to testing independence — form the foundation for data science, machine learning, and evidence-based decision making. Keep practising, stay curious, and let the data guide you.

Best of luck with your exams and your data science journey!

BDS113 Introduction to Statistics · B.Sc. Data Science, Semester II · All 40 Lectures Complete