- Spacy-model for word embedding
- Convert article content into word vectors
- https://github.com/explosion/spacy-models/releases//tag/en_core_web_lg-2.2.5
- Installation
pip install spacy
python -m spacy download en_core_web_lg
- Embedding Neural Network for content features:
- tensorflow
- keras
- Input Data (cleaned data, no NaNs and deduplicated):
- Article contents: articleId (string), title (string), category (string)
- User behavior: UserId (string), articleId (string)
Five arguements:
- two input filenames: TEST_ARTICLES.csv (articleId, title, category) and TEST_CLICKS.csv (userId, articleId)
- NUM_ARTICLES: number of articles selected from
- TOP_NUM_ARTICLES: select NUM_ARTICLEs from the top popular articles
- NUM_EVAL_ARTICLES: number of articles for evaluation
- K: top-K recommendations
- SEED: random seed
python Comparison_results_companydata.py TEST_ARTICLES.csv TEST_CLICKS.csv NUM_ARTICLES TOP_NUM_ARTICLES NUM_EVAL_ARTICLES K SEED