Household Consumption Expenditure Survey (HCES)
Steps
(1) Train a Linear Regression Model (Data: based on MoSPI’s HCES, NSS 68th Round, 2012 link)
(2) Guard against model-overfitting by applying 2 regularizations: LASSO, Ridge
(3) Use cross validation to systematically optimize "alpha"
(4) Make predictions i.e., predict 'total_expenditure' based on'household_size', 'food_expenditure', 'non_food_expenditure', 'income', 'urban_rural'.
(5) Plot co-relation heatmap, the coefficients, and actual vs predicted values