Q1.
Conceptual Foundations (Theory – Hypothesis Testing)
a) Null and Alternative Hypothesis
● Null Hypothesis (H₀): No effect / no difference
● Alternative Hypothesis (H₁): There is an effect / difference
Example:
H₀: Average waiting time = 20 min
H₁: Average waiting time ≠ 20 min
b) Type I and Type II Errors
● Type I Error (α): Rejecting true H₀ (False positive)
● Type II Error (β): Accepting false H₀ (False negative)
Example:
● Type I: Concluding system is slow when it’s actually normal
● Type II: Saying system is normal when it's actually slow
c) One-tailed vs Two-tailed Tests
● One-tailed: Direction specific
Example: mean > 20
● Two-tailed: Any difference
Example: mean ≠ 20
d) p-value and Level of Significance
● p-value: Probability of getting result assuming H₀ is true
● α (level of significance): Threshold (commonly 0.05)
Decision rule:
If p ≤ α → Reject H₀
If p > α → Accept H₀
Q2. Hypothesis Testing – Numerica
Given:
n = 36, mean = 22, σ ≈ 6, μ₀ = 20, α = 0.05
a) Hypotheses
H₀: μ = 20
H₁: μ ≠ 20
b) Test (Z-test)
Critical value at α = 0.05 (two-tailed) = ±1.96
c) Interpretation
Z = 2 > 1.96
Reject H₀
d) SaaS Platform Method
Use tools like:
● Excel / Google Sheets
● SPSS / R / Python
Steps:
1. Enter sample data
2. Use formula or built-in test ([Link] or [Link])
3. Get p-value
4. Compare with α
Q3. Chi-Square & ANOVA
a) Chi-Square Test
Hypotheses
● H₀: Study hours and result are independent
● H₁: They are dependent
Step 1: Observed Table
Hours Pass Fail
<5 12 18
5–8 25 10
>8 30 5
Step 2: Test Statistic
After calculation:
χ² ≈ 20.1
df = (3−1)(2−1) = 2
Critical value (0.05) = 5.99
Interpretation
20.1 > 5.99 → Reject H₀
Study hours and passing are dependent
b) When ANOVA is preferred
● Use ANOVA when comparing more than 2 groups
Example:
Comparing marks of 3 classes
t-test only works for 2 groups
Q4. Time Series Components
a) Components
1. Trend: Long-term increase/decrease
2. Seasonality: Regular pattern (monthly/yearly)
3. Cyclic: Business cycles (years)
4. Irregular: Random fluctuations
b) Examples
● Trend → Increasing smartphone sales
● Seasonality → Ice cream sales in summer
● Cyclic → Economic boom/recession
● Irregular → Pandemic impact
●
c) Importance
● Improves forecasting accuracy
● Helps understand patterns
● Avoids misleading predictions
Q5. Time Series Forecasting
Data
120, 135, 150, 160, 170
a) 3-Month Moving Average
● Feb–Apr = (120+135+150)/3 = 135
● Mar–May = (135+150+160)/3 = 148.3
● Apr–Jun = (150+160+170)/3 = 160
b) Forecast for June
● 160 units
c) Exponential Smoothing (α = 0.3)
Formula:
Assume initial forecast = 120
Stepwise:
● Feb = 120
● Mar = 124.5
● Apr = 131.85
● May = 140.3
● June ≈ 149.2
●
d) Comparison
Method Forecast Nature
Moving Avg 160 Smooth, lagging
Exp Smoothing 149.2 Responsive
Q6. Case Study
a) Hypothesis
● H₀: Sales before = after
● H₁: Sales increased
b) Test
Paired t-test (same store before/after)
c) Time Series Components
● Trend → growth after pricing
● Seasonality → festive sales
● Cyclic → economic trends
● Irregular → sudden demand spikes
d) Forecasting Method
● Exponential smoothing → short-term
● ARIMA → complex patterns
Recommend:
● Simple data → Exponential
● Complex data → ARIMA
e) SaaS Tools
Use:
● Excel / Google Sheets
● Power BI / Tableau
● Python (Colab)
For Hypothesis Testing:
● Built-in statistical tests
For Forecasting:
● Time series models
● Auto forecasting tools