This data is about whether or not different consumers made a purchase in response to a test mailing of a certain catalog and, in case of a purchase, how much money each consumer spent. We built predictive models to predict how much will the customers spend; Spending is the target variable (numeric value: amount spent).
built numeric prediction models that predict Spending. Used linear regression, k-NN, and regression tree techniques. 10 fold cross validation was used on each of the models.
Feature engineering (i.e., create new features based on existing features) was performed to optimize the performance of linear regression, k-NN, and regression tree techniques. Hyper parameter tuning was further performed to further optimize the results.