Q2 Logistic Regression model (Use trainData for this questio…
Q2 Logistic Regression model (Use trainData for this question) (30 points) ( 5 points) a) Create a logistic regression model using Completed as the response variable and the following predicitng variables: Hours_Studied, Age, Region. Call it *model1*. Display the summary. Using “model1”, interpret the coefficients of the following predictors below with respect to BOTH the log-odds of completed and the odds of completed. 1) Hours_Studied 2) RegionWest (5 points) b) Using the “trainData” dataset, create a logistic regression model using Completed as response variable and all variables in “trainData” as predictors (call it model2) and display the summary of model2. Using the model coefficients, how would you compute the predicted probability of completion for a 30-year-old male from the South enrolled in the free course who studied 10 hours and submitted 5 assignments? (10 points) c) This dataset is without replications. Aggregate the response data to convert it into binary data with replications. (Use categorical variables only)Fit the following logistic models:i) Using aggregated data. Call it model.agg. Display the summaryii) Using the non-aggregated data. Call it model.withoutagg. Display the summary. (Use the same categorical variables as in model.agg) What is the difference between the coefficients of both the models? ExplainHow are the null and residual deviance different for both models? Explain (3 points) d) Perform a test for overall regression of the logistic regression “model.agg”, using a significance level of alpha=0.05. Does the overall regression have explanatory power? Provide interpretation of the test. (3 points) e) Perform the goodness of fit test on model.agg using deviance and Pearson residuals and interpret the results. (4 points) f) Evaluate the predicting power of the “model.withoutagg” in Q2c using cross validation. In your evaluation, analyze how different cutoff thresholds affect the model’s prediction error