A comparative analysis of data preparation algorithms for customer churn prediction

摘要: Data preparation is a process that aims to convert independent (categorical and continuous) variables into form appropriate for further analysis. We examine data-preparation alternatives enhance the prediction performance commonly-used logit model. This study, conducted in churn modeling context, benchmarks an optimized model against eight state-of-the-art data mining techniques use standard input data, including real-world cross-sectional from large European telecommunication provider. The results lead following conclusions. (i) Analysts better acknowledge technique they choose actually affects performance; we find improvements of up 14.5% area under receiving operating characteristics curve 34% top decile lift. (ii) enhanced logistic regression also competitive with more advanced single ensemble algorithms. article concludes some managerial implications suggestions research, evidence generalizability other business settings. study impact on customer performance.Effective improves AUC lift 34%.Optimized