This session consist of two parts. Both parts include the use of Wordstat text analytics in combination with Minitab time series tools and predictive analytics.

In part one of the meeting the presenters reveal the results of a yearlong time series study of text extractions from a balanced, consistent sample of daily news headlines published September 2022 through August 2023. The presenters demonstrate how to use headline text to build forecast models for predicting DOW and S & P performance. These examples illustrate a general method for using time-based text analysis to improve business forecasting by combining sentiment and news topic modeling with the more traditional economic predictors such as the Fed Fund Rate and the Volatility index (VIX). The presentation includes a discussion of how machine learning tools aid topic selection and greatly facilitate the forecast model building process.

Part two of the meeting demonstrates the use of topic modeling combined with machine learning to aid business decision making. The presenters demonstrate a use case where textual phrase extractions from samples of public financial documents (in this case corporate prospectus documents) can be used to train machine learning classification models. The models are then applied to make predictions that aid corporate value assessments for mergers, acquisitions, and stock buying/selling.