(57m) Text Mining and Analytics (poster)

Text mining seeks to extract information and insights from textual data. In particular the distillation and summary of documents is of increasing commercial interest, from news media to patents to maintenance logs.

Application areas include business analysis, scientific discovery, predictive maintenance, marketing and communications, litigation, and cybersecurity, among others.

Basic workflow elements will be reviewed in this talk, including: definition of the corpus (which may involve the pre-step of converting images to text), preprocessing, visualization, filtering, statistical analysis, word embeddings, and classification.

Some more advanced subjects, for example topic modeling and sentiment analysis, will be introduced and described.

Demonstrations of functionality will be presented using MATLAB.