Meeting 28

Title:Optimization on K-modes

Date: 05/04/2020

Time: 5:30 pm     

Location: Zoom Conference Call

Attendees: Sofian Ghazali, Rahul Balamurugan, Zahid Kamil

Absences: None

 

Action Items: Optimizing K--modes and GUI

Meeting Agenda:

1. Any steps to optimize K-modes

2. GUI integration

Discussion and Notes:

The meeting was focused on determining if our K-modes is sufficient and we decided to move on with K-means analysis. We know that K-means would not require as much preprocessing as the K-modes algorithm. Based on previous experience we can finish the K-means as fast as we could. The next step is how to integrate a user's input to be compared with the existing data and detect an anomaly. We later learned that simple tasks such as calculating the distance between two matrices or values or tables or whatever you call it would be enough to determine an anomaly. This is very ture because at the end we ended up with clustering centroids in a csv file and the users inputs both converted to numpy arrays to calculate a distance score. This was the same that was done with the anomalous data and so this idea seeming integrates with the GUI and we obtained good results.

The next meeting will be with Dr. Boutros and we are looking forward for his input into our long and strenuous work. Hopefully, he is proud of our work since we are almost at the end of senior design. The next step would be to finalize implementing K-means.

Next meeting action items:​

1. Meeting with Dr. Boutros,

2. Finalize K-means

Meeting Adjournment: 6:45 pm

Meeting Minutes prepared by: Zahid Kamil