The Basics
A very basic approach to modeling of text in JavaScript.
Vectorspace
Stopwords
You can upload a text file of newline separated stop words. These words will be stripped from the input documents.
Queries
Need to allow uploading loads of files for querying.
Enter query:
Document Comparison
You can enter text into each box and compare the two documents.
Document 1
Document 2
Bigrams
- Enter text into the box.
- Click "learn" for it to update the model
- Click "write" to generate words, or type directly into the box.
Presently it always tries to build the highest probability pair (although given equal probabilities it randomly chooses). I'd like to have it optionally choose randomly against a weighted pool so it's more likely to choose that path, but won't every time.
Trigrams
String Cleaning
Just for testing what happens when strings are sent in.