Team Houdini

Netflix_Recommender-Collaborative_Filtering -- Yehuda Koren

How to clone this repo

git clone https://github.com/adi-iiith/Netflix_Recommender-Collaborative_Filtering.git

Requirements :

• Python version - 3
• sparse matrix library - scipy.sparse

Objectives :

• Implement the paper, reduce the RMSE error for movie prediction and perform better than the currently existing models ( Latent Factor and Nieighbourhood ).

Dataset

• Use dataset as " ../input/u.data "

Results :

• Movie previously watched by user.

• Movies predicted.

• For further explanation refer to Team Houdini.pptx or Team Houdini.pdf

How to Run the project :

• Save the Dataset in "../input" folder.

• (python v3.x)

• To train and test Integrated Model :

  --Train model -- " python3 integrated_model.py " --Test model -- " python3 integrated_train.py " ( OR ) -- Train and Test (per epoch ) -- " python3 integrated_model_epoch_error.py"

• To train and test Neighborhood model :

  • Train and Test (per epoch ) -- " python3 neighborhood.py"

• To train and test SVDpp model :

  • Train and Test (per epoch ) -- "python3 svdpp.py"

• Predictions for Integrated model (Demo)

  • test_beta.ipynp

• EDA for Dataset :

  • EDA.ipynp

.. Project Report

.. Project Presentation