BigDataAlgGroup4

Movie recommendation engine using the content-based filtering and collaborative filtering

1. install anaconda 3.X

2. create virtual environment

conda create --name myenv
conda activate myenv  

3. install reqiured packages listed in the code/requirements.txt

conda install --file requirements.txt  

4. setup of the folders for scripts and data

scripts should be under the code folder
datasets should be under the data folder
models should be stored under the model folder

5. to run content-based filtering, call content_based_gc.py and specify parameters: -tl, -lsh, -st e.g.

-tl: title of the query movie, str
-lsh: whether to use LSH to put similar movies in same bucket first, str, y or n
-st: if use LSH, the top movies should be sort on popularity pop or cosine similarity cosine

python content_based_gc.py -tl MovieTitle -lsh n   
python content_based_gc.py -tl MovieTitle -lsh y -st cosine   

6. to run collaborative filtering, call collab_model_SVD_gc.py and specify parameters: -uid, -iid, r_ui, -fn. e.g.

-uid: user ID, interger
-iid: movie ID, interger (optional)
-r_ui: real rating for the given user-movie pair, float, (optional)
-fn: filename to save the trained svd model or to import the existing svd model, str

python collab_model_SVD_gc.py -uid 1 -iid 31 r_ui 2.5 -fn model_filename  
python collab_model_SVD_gc.py -uid 1 -fn model_filename