Building a response model based on the Portuguese bank marketing data set from UCI
- Acquisition+Analytics+-+Assignment+-+Solution : Main ipynb file with code and solution
- bank_marketing : Dataset
- Data+Dictonary+-+Bank+Marketing : Data dictionary for dataset
- Recommendation : Insights found fom the solution.
Based on the bank marketing data set you need to build a response model.
You wanted to predict the probability of a response from each prospect and target the ones most likely to respond to the next telemarketing campaign. The steps were as follows:
- Identify relevant predictor variables for a response using EDA.
- Build predictive models and choose the best one.
- Sort the prospects in order of decreasing probability of response (predicted by the best model) and target the top X% (or top Y deciles), where X would be determined by your business objective (e.g., maximising the overall response rate/number of responders at a fixed marketing cost).
You should resolve to build another model without including the variable ‘duration’.
Also, set the business objective to achieving 80% of total responders at the minimum possible cost. The total number of responders is the total number of prospects who responded, from the available data of about 45,000 prospects.
Based on this information, calculate the X in the top X%, i.e., how many prospects should be called to meet the business objective.