Predicting Customer behavior is key to marketing strategies. Customer Relationship Management technology plays a very important role in business performance. Predicting customer behavior enables the business to better address their customers and enhance service level and overall profit. A model based on Support Vector Machines is proposed used to classify clients and predict their purchases in a real retail department store. Different Kernels functions are used and their performance is evaluated. The data scaling is implemented with SVM model and its performance is evaluated. The study is conducted in a real retail department store in Albania for the 2020 year. Implementations in python shows that the proposed model performs better in time and accuracy.

Keywords : Classification; Cross-Validation; Machine Learning; Neural Networks; SVM, Standard Scale Function.