DATA SCIENCE METHODS USED TO STUDY HUMAN BEHAVIOR
My research with Professor Sudeep Bhatia was on data science methods used to study human behavior. It mainly entailed an exploration of computational modeling on binary decision making with very short reaction time. The model we explored is called the Drift Diffusion Model and it is a model that describes the decision processes in risky choice scenarios. It suggests that the deliberation process in binary choice can be described by a stochastic dynamic variable which is either moving toward acceptance or toward rejection. The DDM predicts that the starting point of the process determines the amount of time taken to decide, as well as the likelihood of accepting and rejecting. Through experiments we conducted in the lab, measuring reaction time and computational modeling, we investigated how changing the starting point influences the decision made and the time that is taken to reach the decision. We were able to confirm that in a binary decision like a gamble, the time taken to accept the decision depends on the starting point of the drift and the combined drift rate. In addition to developing data science skills applicable to the study of human behavior, I also spent time reading the work related to the DDM and decision making. Through this broad experience, I was able to learn that the knowledge of the field is as important as acquiring the data science skills in order to make sound improvement in the work as there comes a point where everyone in the team has all the fundamental analytical skills. However, I managed to explore and learn more data science methods used in the field of psychology like natural language processing (NLP) and queering online knowledge bases like concept net which are very important for social sciences. Last but not least, research has taught me how to navigate academic spaces and strive to take initiative, be creative and take the lead all thanks to Professor Sudeep Bhatia.