ml-classify-text-js
ml-classify-text-js copied to clipboard
andreekeberg
Machine learning based text classification in JavaScript using n-grams and cosine similarity
`tokens` is an object with key of words from input, so if one of words is `constructor`, then check fails. https://github.com/andreekeberg/ml-classify-text-js/blob/d8b18d2c17a6d03170bb65848d26ba6725bd4dfe/src/classifier.js#L223
Hi, I was trying to install module on linux server, was able to install the same on mac. `npm WARN EBADENGINE } npm ERR! code 127 npm ERR! path /home/www/test/node_modules/core-js-pure...
Add support for for passing custom methods to `Classifier` and `Model` for defining custom getter/setter methods that allows us to bypass the `data` object literal that is currently stored directly...
The original `tokenize` method assembled a `sequence` of words using the space character as a delimiter, then counted tokens using `split()` on the space character. This fails if there are...
When making predictions, and a vocabulary is used, the complete vocabulary (with any new terms not already in the current model vocabulary) needs to exist as a separate copy that...
Need to implement a smarter method of tokenization which takes into account languages that traditionally does not use spaces between words (currently resulting in full-sentence tokens not suitable for the...
Add an easy to use command line tool to easily train models from large amounts of files, preferably in a format like the [BBC datasets](http://mlg.ucd.ie/datasets/bbc.html).
Create some practical examples and add these to the documentation, like sentiment analysis, profanity detection, etc.
Add a simple `StopWords` class and a related `stopWords` getter and setter in the `Model` class, which will store a list of words in a `Set`, and remove these both...
Build upgrade, ESM compatibility, using biome, tsup and vitest.
Metadata
Owner
Metadata
Machine learning based text classification in JavaScript using n-grams and cosine similarity