Client Ranking System
Rank clients according to the amount of attention needed to be given to them.
Data were extracted from our CRM implemented with Dynamic365 and ERP implemented with BC systems. Different relevant features were computed for each client from these data. After that, data for each client was labeled in one of the five classes (lowest, low, moderate, high, and highest attention). Then, an ordinal classifier was used for the prediction.
Technologies used: Scikit-learn, xgboost, Dynamic365, Numpy and Pandas