Perceiving Personality through Sentiment Analysis

Abstract

The goal of this project is to find new ways to understand the human personality through the analysis of writing style over long periods of time. Techniques from Natural Language Processing and Sentiment Analysis are used to perform the analysis. The data used for analysis consists of large sets of personal letters written by John Adams, Thomas Jefferson, and George Washington - these three figures were chosen because they all have large, publicly available collections of letters from the Massachusetts Historical Society and the University of Virginia Press, and they were all forced to deal with a similar set of events in their lives. We attempt to analyze the overall writing styles of the presidents and how they change over time. In addition, we try to identify the major events of their life and analyze the ways that these presidents write about them. This analysis allows us to pinpoint differences in the ways that significant life events are perceived, what emotions they trigger, and how each person responds to them; thus, we find new ways to understand the human thought process and how it differs from individual to individual.