Logistic Regression in Python – Case Study

Consider a bank that asks you to develop a machine learning application to help them identify potential customers who will open a term deposit (some banks also call it a time deposit) with them. The bank regularly conducts surveys, either by phone or through web forms, to collect information about potential customers. The surveys are general in nature, reaching a very large number of respondents, many of whom may not be interested in doing business with the bank. Of the remaining respondents, only a few may be interested in opening a time deposit. Others may be interested in other facilities offered by the bank. Therefore, the surveys are not necessarily designed to identify customers who will open a time deposit. Your task is to identify all customers who have a high probability of opening a time deposit from the large amount of survey data that the bank will share with you.

Fortunately, for those interested in developing machine learning models, there is publicly available data. This data was prepared by students at the University of California, Irvine with external funding. This database, part of the UCI Machine Learning Repository, is widely used by students, educators, and researchers around the world. This data can be downloaded from here.

In the following sections, let’s use the same data for application development.