Fuzzy particle swarm optimization (FPSO) based feature selection and hybrid kernel distance based possibilistic fuzzy local information C-means (HKD-PFLICM) clustering for churn prediction in telecom industry

Abstract Customer churn has been considered as one of the key issues in operations corporate business sector, it influences turnover directly. In particular, telecom industries are seeking to develop new approaches predict potential customer churn. So, needs appropriate algorithms overcome increasing problem This work proposed a prediction model that employs both strategies classification and clustering, helps recognizing consumers giving reasons after churning subscribers industry telecom. The process information gain fuzzy particle swarm optimization (FPSO) executed by method feature selection, besides divergence kernel-based support vector machine (DKSVM) classifier is employed categorizing customers approach. this way, compelling guidelines on retention have generated since plays vital role relationship management (CRM) suppress churners. After process, divided into clusters through fragmenting data customer. cluster-based offers provided clustering algorithm hybrid kernel distance-based possibilistic local C-means (HKD-PFLICM), whereas measurement distance accomplished functions such hyperbolic tangent Gaussian kernel. results reveal (FPSO- DKSVM) produced better compared other existing K-means, flexible K-Medoids, (FLICM) , FLICM (PFLICM) entropy weighting (EWFLICM). Article highlights major concern most companies performance improved applying artificial intelligence learning techniques. Churn crucial industry, they position maintain their precious organize Relationship Management.