PadhaikoAll subjects

Past exam question: 2081 Chaitra (Regular), Q6(b)

Foundation of Data Science · 5 marks

Open in the Padhaiko app

Topic: Empirical Models, Simple Linear Regression, MLE & Least Squares

Question

What is regression, and how does it support predictive analytics? Explain with a small sample business application.

Solution

Definition and Mathematical Formulation of Regression: Regression analysis is a supervised statistical learning technique used to model and analyze the relationship between a continuous dependent (target/response) variable and one or more independent (explanatory/predictor) variables .

In Simple Linear Regression with a single predictor:

where:

Model parameters are estimated via Ordinary Least Squares (OLS), which minimizes the Residual Sum of Squares (RSS):

How Regression Supports Predictive Analytics:

  1. Quantitative Forecasting: Evaluates fitted equation to project future outcomes for novel, unseen input values.
  2. Quantifying Marginal Effect and Sensitivity: The magnitude and sign of reveal whether a variable has positive or negative leverage and quantify the exact expected return per unit input.
  3. Hypothesis Testing and Significance: Uses -statistics and -values to evaluate whether observed associations are statistically genuine or mere random noise.
  4. Scenario Simulation ("What-If" Analysis): Enables decision-makers to simulate varying operational budgets, pricing strategies, or market conditions to forecast risk and optimal resource allocation.

Small Business Application: Advertising Budget vs. Product Sales: A retail enterprise wants to predict monthly product revenue (in thousands of dollars) based on digital advertising expenditure (in thousands of dollars). Data collected over 5 months:

MonthAd Spend ($k)Revenue ($k)
1215-4-114416
2421-2-5104
36260000
4832+2+6124
51036+4+104016
Sum3013010640
  1. Calculate Sample Means:
  1. Compute OLS Estimators:
  1. Predictive Equation:

Business interpretation: Baseline monthly revenue without advertising is . For each additional invested in digital advertising, monthly revenue increases by an expected .

  1. Predictive Application (Forecasting): If management plans to allocate () for next month's advertising campaign:
Answer:

Regression models the relationship between a continuous dependent variable and one or more independent variables via Ordinary Least Squares. It supports predictive analytics by estimating parameters that enable quantitative forecasting and sensitivity analysis. In the business example, fitting ad spend to sales yields , allowing the firm to forecast revenue of for a planned advertising expenditure of .

All past exam questions for Foundation of Data Science