Week in Insights: Text Classifiers Are Tax-Tracking Trailblazers

Tracking federal tax laws can be an exercise in endurance. The challenge is even greater at the state level, given the countless draft bills and revisions across 50 jurisdictions.

But what if machine learning could help track those legislative changes, categorize tax laws, and predict policy trends long term? Researchers in neural linguistic programming have now achieved this. They used machine learning to train text classifiers to extract tax law changes from a century’s worth of US state legislative proceedings.

Text classifiers use keywords and phrases to classify documents. This makes it easier to understand why a piece ...