eloukas / job-recommendation-system

A lightweight demo app which recommends similar jobs. Combines unsupervised ML + IR.

job-recommendation-system

Get recommendations of similar jobs when job hunting.

The system uses unsupervised learning to generate the recommendations. More specifically, the pipeline includes:

• Extensive data preprocessing pipeline (HTML/URL/stopwords removal, lowercasing, stemming)
• Integer Encoding
• TF-IDF text vectorization
• SVD dimensionality reduction
• Cosine similarity metric

Example

Install

• Before starting, ideally, it's recommended to switch to a virtual environment first via conda, using Python 3.8.
• Install dependencies in your virtual environment via pip install -r requirements.txt

Run

• To train or inference the model, run python run.py --dataset_path <DATASET_PATH> --mode <MODE>. The <MODE> parameter should be between train, dev, test.
• To run the demo, run streamlit run demo.py.