Current Clustering algorithms that we have, provide very less insight regarding cluster membership and provide very less interpretability. Due to this, we are not able to explain the recommendations made by the algorithm, whereas in the industry we often need a detailed explanation regarding the recommendations. In this project, we have tried to solve this problem by introducing a new clustering technique. This increases the interpretability of clusters forms. We will be leveraging the rules generated by the Decision Trees which forms the basis of our clustering algorithm. By doing this we provide a more clear understanding of why the specific data point belongs to the respective cluster by giving insights on every rule used.
Read more about results here Paper
Main base code resusable for any dataset BaseCode
You can always pull issues or mail us for any issues or querries.
Have a nice day!