Rule mining and prediction using the flek machine – A new machine learning engine

Abstract

One of the exciting areas in data science is the development of new machine learning engines to do data mining, analytics and prediction. In this paper, we introduce the “Flek Machine” – an innovative AI engine that learns a Bayes Net model from binary data. FlekML, the core machine learning engine inside, builds a rich model that can be manipulated by the Toolkit to do rule mining, discover associations and association maps as well as make predictions. The Flek Machine enables binary, multi-class and multi-label classifications all on the fly and over the same built model. This tool has several use cases such as customer behaviour analysis, predicting equipment failure in IoT, or detecting drug combinations that produce side effects.