The Power of Sentiment Extremes in Predictive Analytics

This meeting will be presented in two parts. The first segment is titled: Utilizing Machine Learning & Neural Networks for Company Valuation and the second part of the presentation will cover The Power of Crowd Behavior in Predictive Analytics.

Parallax Financial Research CEO Kris Kaufman will explore two applications of advanced quantitative analysis covering both the search for value and the measurement of market behavior.

The Parallax method of searching for value is centered around a suite of industry-scoped multi-factor neural network equity models that were pre‐trained to convert recently reported company fundamentals into a current estimated market price. This algorithm automatically performs an appraisal by simultaneously using common stores of value like earnings, cash flow, book value, and dividends, along with critical modifiers such as industry, margins, debt, interest rates, and inflation, to determine a price appraisal.

The key advantage of using this method is managers can utilize the industry and sector specific interrelationships between fundamental factors that are not obvious by using classic ratio ranking methods. Energy companies have very different responses to commodity prices than do utility companies. Interest rate changes can affect valuations in different industry groups quite dissimilarly. Additionally, by removing the biases and precognitions that often come from human-made valuations we can better assess the true value of a company given the current market environment.

The second part of this session will focus on predictive analytics developed through a model called ExtremeHurst. ExtremeHurst is a quantitative detector of extreme investor behavior that signals the beginning or end of a trend. Strong trend-persistent stock price movements are evidence of positive feedback (i.e., investors buying because the price is rising, driving prices higher), while extremes of mean reversion are evidence of negative feedback. ExtremeHurst exploits the science of non-linear dynamics to identify unique and predictive signals occurring in freely trading auction markets.