Interpreting Word2vec/GloVe word embeddings

In this repository I attempt to reproduce the findings of a paper written by Thomas A. Trost and Dietrich Klakow titled Parameter Free Hierarchical Graph-Based Clustering for Analyzing Continuous Word Embeddings.

I use scikit-learn, Neo4j, and the Neo4j graph algorithms and APOC libraries.

Installation

• Download the Neo4j Desktop from neo4j.com/download

• Install the APOC and Graph Algorithms libraries

• Install Python required libraries

pip install -r requirements.txt

Running the algorithm

python glove_to_neo4j.py
python glove_nn.py