Understanding customer response on telemarketing campaign and finding customer profile for more campaigns. We can see which customer will deposit money and likely to use more campaigns so we will focus on that customer.
We need a type of data which can tell about the customer , weather he/she has answered the calls , has account, diuration talked etc.
Dealt with imbalanced data hence, prediction includes bianess
Customers who are married have high chance of positive response on bank call about schemes. Customers gave more response in month of march and september. Customers response also depend on duration of call Customers response depend on the age range too.
Decision tree model is working better than KNN model We can drop variables which are not contributing towards the model performance. We should consider f1 score rather than model accuracy.