This jupyter notebook constains a machine learning study about The CoIL Challenge 2000. This event was a competition to prefrom an analisys to find future clients of an insurance company. That company wants to make an intelligence campaing in order to generate an incrementation of its revenue.
The aim of that competition and dataset is to find the people who become clients and separate for those who dosen't want to contract this service.
From machine learning point of view, this dataset is unbalanced and challenging due to the 'unusefull' data which it constains.
It is a large study about different machine learning algorithms, from logistic regression to 'deep' learning approach. To solve the problem of inbalance different techniques are taking into account: oversamplig minority class (SMOTE), undersampling mayority class and make the cost policy inbalanced(SVM and Deep Learning).
The proyect was developed in Python with the use of Keras and Sklearn packages.