ML-science kit

Simple Data Science and Machine Learning toolkit

Personal revamped version of KelosBot classifier's libraries

What it could be..?

Here you will find out scripts for:

• Make VK dataset
• Perform EDA with csv datasets with cleaning texts
• Train Keras model, which we`ve used in KeLoss project and save it
• Classificate natural-language texts
• Estimate the semantic similarity of pairs of texts

Something about parameters..?

Yeah, it could be interesting. Python Classes stored in these files uses very much parameters.
The library is well documented. More details about its use can be learned using the "help" method for each class.

Prerequireties

• Keras
• Tensorflow
• pymorphy2 with pymorphy2-dicts-ru/uk
• nltk with basic dictionaries and stop-words
• numpy
• pandas

If you want to use our dataset

• vk_api

Then you could make dataset, possibly with our scripts, stored in this branch

Now you can train your own model

And after all, PREDICT THEM ALL!!!