*We are provided with over a hundred variables describing attributes of life insurance applicants. *The task is to predict the "Response" variable for each Id in the test set. "Response" is an ordinal measure of risk that has 8 levels. *Handling null values and the class imbalance. *Built tree based models,logistic regression and stacked models. *Doing a feature importance check for all the models. And then, interpreting the model, how every feature is pushing towards which target response.
#Data Wrangling And Visualizations Null values found, filling missing values using appropriate functions. Changing targer variable to lesser categories. Plotting heatmap to find the correlation between continuous features. Plotting displots, barplots and boxplots for the features.
#Model Building I took a split on the data with training data as 80% and test data as 20%. I took a split on the data with training data as 80% and test data as 20%. Built base level models for Logistic Regression, Random Forest,Gradient Boosting, Voting Classifier and stacked model. After building base model, I tuned the hyperparameters using grid search with 2 cross validations. Scoring metric used ROC-AUC.
#Model performance
Gradient Boosting, Voting Classifier and Stacked models are performing really well. Their train and test errors and also the roc scores and f scores are really close and good.