Does it worth to devote time on this customer?

Kaggle: Santander Value Prediction Challenge

Project

Abstract

To stay competitive in the banking sector, the banks have to provide more special services from customer to customer. The data is taken from Kaggle challenge of Santander Group who aims using these predicted results in order to give an improved service consisting of personalization and customization. In order to conduct prediction of the value of customer transactions; neural networks, support vector regression and Gaussian processes are utilized. Following the several trials, runs and parameter tuning, the performances of the best models of each method are validated and compared as a final step. The future research directions are suggested as follow up procedures.

Keywords: Gaussian processes, neural networks, prediction, support vector regression
Return to projects page