Parallel Programming of Churn Prediction Using Gaussian Naïve Bayes

This paper presents churn predictions with the Gaussian Naïve Bayes method. Churn prediction is a forecasting method to predict customer decisions in a company's service or product (churn). With high public enthusiasm and an increasing number of customers in the Big Data era, a fast computing process is needed to predict churn as quickly as possible. In this paper, computing is accelerated by the OpenMP platform parallel algorithm. Churn prediction experiments are performed with different amounts of test data, ranging from 100, 300, 500, 700, to 900 data. The results obtained show that implementing OpenMP in predicting churn is faster than serial processing. The obtained speedup and efficiency reached more than 1.49 and 37%, even for test data of 300 and 500, based on the tests, the speedup and efficiency reached 1.99 and 50%.