Sentiment Analysis Parser Demo



Training models used



  1. Binary - Netflix Review Model: a model trained by examining film and TV show reviews on Netflix.
  2. Binary - HT Provider Review Model: a model trained by examining services reviews.
  3. Categorical - Stanford Model: a model trained by examining data provided by Stanford University.
  4. Categorical - HT Provider Review Model: a model trained by examining HT services reviews.


To find out more about the models we trained please visit our Models and Datasets pages.





Datasets evaluated



  1. Gun Ads: the MEMEX gun ads data. The gun ads/reviews get examined for their sentiment.
  2. HT Reviews: the MEMEX services reviews examined for their sentiment and therefore possible issues.




Running the parser on 100 gun ads



Binary Sentiment Analysis Parser



  1. Overall indication of the sentiment




  2. Average distribution of sentiment among countries




  3. Average distribution of sentiment among states






Categorical Sentiment Analysis Parser



  1. Overall indication of the sentiment






Runnig the parser on ALL gun ads



Total # of ads: 3,929,469



Categorical Sentiment Analysis Parser



  1. Overall indication of the sentiment






HT ads sentiment distribution



BINARY



  1. Netflix model results




  2. Binary HT Provider Review model results






CATEGORICAL



  1. Stanford model results




  2. Categorical HT Provider Review model results






HT ads sentiment distribution - HT Negative Data (new)



BINARY



  1. Netflix model results




  2. Binary HT Provider Review model results






CATEGORICAL



  1. Stanford model results




  2. Categorical HT Provider Review model results