Telecommunication company big data classification by Data Mining technique

This work deals with analysis of telecommunication company customer information by data mining technique. For this purpose, we used cluster analysis method of k-means and support vector machines. The first was applied with non-conventional k-means++ centroid initialization method. Special software on C# language was developed that transforms initial data of telecommunication company to comma-separated values. We designed software on Python language with library sklearn for data clasterization. Data processing was performed, and customers with “aberrant behavior” were detected. Telecommunication company should make decision for those customers by itself. We developed classifier learned on big data based on support vector machines to classify new data of company.