Back

Speaker "SK Reddy" Details Back

 

Topic

 
How to solve the curse of imbalanced data

Abstract

It is a rare phenomenon to obtain a balanced dataset. Modern industrial and enterprise datasets are notoriously imbalanced. There are many reasons to this state of affairs. Cost of collecting data and incomplete feature awareness are some. But, many times, data for certain class does not exist. For example in Credit Card Fraud Detection or in Medical Diagnosis, it is difficult to find equal number of Positive and Negative data points.
 
I would like to discuss various practical scenarios on why imbalanced data, what techniques to use to address the data and its impact on how the model behaves.
Attendees: Beginners and Intermediate skilled in machine learning.
Prerequisite knowledge: Preliminary understanding of classical machine learning techniques.
What you'll learn: You would learn various statistical and ML techniques to process data, adjust models and re-learn interpretation of model predictions.

Profile

SK is the Chief Product Officer AI in Hexagon (www.hexagon.com). He is an AI expert and a successful twice startup entrepreneur. He is an AI startup advisor too. Also he is a frequent speaker in conferences and is an AI blogger. LinkedIn: www.linkedin.com/in/sk-reddy/ Blogs: www.linkedin.com/in/sk-reddy/detail/recent-activity/posts/ YouTube Channel: https://www.youtube.com/user/skreddy99/videos