Natural Language Processing - Deciphering the Message Within the Message Stock Selection

Explore natural language processing techniques for stock selection using earnings call transcripts in this 22-minute video from Open Data Science. Learn how to decipher sentiment- and behavioral-based signals that have demonstrated historical stock selection power in the U.S. market. Discover the ABCs of NLP, understand its importance in finance, and delve into the general steps of NLP analysis. Examine the process of text preprocessing, signal construction for both sentiment-based and behavioral-based indicators, and review empirical results. Gain insights into controlling for risk and alpha factors, natural tilts of sentiment-based signals, and correlation analysis. Enhance your understanding of how unstructured data can be leveraged to differentiate sources of alpha in investment strategies.

Intro
The ABCs of NLP
Motivation - Why is NLP Important?
NLP General Steps
Stock Selection Insights Using Earnings Call Transcripts
Motivation - Historical Performance Comparison of Strategies
Earnings Call - Introduction
Text Preprocessing prior to Signal Construction
Two Main Categories of Signals
Construction of Sentiment-Based Signals: Sentiment Level
Construction of Sentiment-Based Signals (continued): Change in Level and Change in Trend
Construction of Behavioral-Based Signals
Description of Empirical Results
Sentiment-Based Signals Empirical Results
Sentiment Based Signals - Entire Transcripts
Controlling for Risk and Alpha Signals
Natural Tilts of Sentiment-Based Signals
Results after control for risk and alpha factors
Correlations
Takeaways - Empirical Results