Document-similarity-using-doc2vec-and-gensim

Python implementation of a document similarity checking using Doc2Vec.

File structure

Document-similarity-using-doc2vec-and-gensim/
├── data/
│   ├── 20news-bydate.tar.gz
│   ├── 20news-bydate-test
│   └── 20news-bydate-train
├── models/
│   ├── doc2vec_model.bin
│   ├── doc2vec_model.model
│   ├── doc2vec_vector.txt
│   └── doc2vec_model.bin.dv.vectors.npy
├── dataset_preprocess.py
├── inference.py
├── README.md
├── requirements.txt
└── train.py

• data/train_data.txt: Training data file
• models/doc2vec_model.bin: Trained Doc2Vec model file
• models/doc2vec_model.bin.dv.vectors.npy: Document vectors file for the trained model
• README.md: Project documentation file
• requirements.txt: Required Python packages
• inference.py: Script to check similarity between two documents
• train.py: Script to train the Doc2Vec model

Installation

Install the dependencies using pip:

gensim==4.2.0
nltk==3.5
numpy==1.23.1
numpy==1.23.2
pandas==1.2.0
scikit_learn==0.23.2

Install the required packages:

pip install -r requirements.txt

Training the Doc2Vec model

python train.py 

Inference

Check the similarity between two documents

python inference.py