Personal data clustering in retail marketing

Abstract

This research proposes a technique for high-dimensional transactional data clustering analysis, which can be used by retail companies to extract meaningful marketing insights from personally identifiable information. These insights help companies to adopt their business processes to changing market conditions and objectively analyze the behavioral patterns of different types of customers. The model uses several data preprocessing algorithms that replace the original feature space with a feature space that exclusively consists of values of irreversible hash functions. The following clustering algorithm processes these hash values in a metric space. Thus, the described approach is compliant with various data regulation laws like GDPR in EU or CCF in Saudi Arabia