Author:
Gevorgyan Ruben,Hakobyan Yenok
Abstract
Blockchain technology allows confidential data to remain strictly confidential and, at the same time, can be used for machine learning with external researchers. Blockchain enables valuable datasets to be reliably processed and speeds up the process of developing valid data mining applications. Blockchain can make it much easier to share datasets, machine learning models, decentralized intelligence, and trustworthy decision-making, which is very important in anomaly detection and fraud detection. This chapter presents a new framework for partitioning categorical data, which does not use the distance measure as a key concept. The matching-based clustering algorithm is designed based on the similarity matrix and a framework for updating the latter using the feature importance criteria. The experimental results show this algorithm can serve as an alternative to existing ones and can be an efficient knowledge discovery tool, especially in anomaly detection using blockchain technologies. While the algorithms for continuous data are relatively well studied in the literature, there are still challenges to address in case of categorical data. Based on the similarity matrix and a novel method for updating it using the feature importance, a matching-based clustering algorithm is designed.