Natural Language
This page contains a set of .air files (and associated python scripts) to help you get up & running with some natural language use cases with AI Squared
Topic Classification
The ag_news_subset data set is drawn from the 4 largest classes of the AG news data set. This model can be used to classify the contents of a webpage as being a member of 1 of these 4 classes (world, science/technology, sports, and business). Try running this over news sites and social media posts!
Sentiment Classification
The IMDB data set is the classic example of positive / negative sentiment classification, having drawn over 25,000 movie reviews from the IMDB movie ratings website. This model will perform sentiment analysis over the contents of a webpage & classify that data as being "positive" or "negative" - note that due to the way the DOM text is harvested and preprocessed may skew your model results.
Last updated