SPLIT AND CONQUER METHOD IN PENALIZED LOGISTICS REGRESSION WITH LASSO (APPLICATION ON CREDIT SCORING DATA)

Fitriyatus Shofiyah; A'yunin Sofro

Open Conference Systems, MISEIC 2018

Fitriyatus Shofiyah, A'yunin Sofro

Last modified: 2018-07-07

Abstract

Big data is one of the biggest issue recently. We need a new approach to deal with the problem. One of a statistical strategy that can be used to solve the problem is split and conquer method. In this paper we focus on non-Gaussian data, i.e binomial distribution. Â We will provide a model of penalized logistic regression with LASSO penalty, how to estimate the parameters and an illustration on real data, i.e credit scoring data.

Keywords

Credit Scoring; Logistic Regression; Penalized Regression; LASSO