aaaksenova / term-paper

Distributional Analysis of Function Words in Russian

Distributional Analysis of Function Words

This repository was created for a term paper. As you can see there are four directories.

1. Theory

   Contains theoretical papers related to the topic.

2. Models

   Contains estimation of various models for word vector representation on test subcorpus.

3. Similarity count

   Contains evaluation of two types of context similarity: maximum pairwise similarity between right context and left context, rigt and left mean context vectors similarity.

4. Main

   Contains code for preprocessing of main train corpus for final CBOW model and discussion of various classification algorithms.