deepaksharmacse12 / Basic-Recommender-System

Basic-Recommender-System

Usage

Collaborative Filtering

1. recommendations.sim_distance(recommendations.critics,'Lisa Rose','Gene Seymour') : gives similarity score between 'Lisa Rose' and 'Gene Seymour'.
2. recommendations.sim_pearson(recommendations.critics,'Lisa Rose','Gene Seymour') : gives Pearson correlation score between 'Lisa Rose' and 'Gene Seymour'.
3. recommendations.topMatches(recommendations.critics,'Toby',n=3, similarity = sim_pearson) : returns the best matches of similar users for 'Toby' from the 'recommendations.critics' dictionary with pearson score as similarity measure and top '3' results.
4. recommendations.getRecommendations(recommendations.critics,'Toby',similarity=recommendations.sim_distance) : get recommendation for 'Toby', not only we get a ranked list of movies, but also a guess at what rating would be.
5. recommendations.transformPrefs(recommendations.critics) : returns dictionary with key values as movie and value as dictionary of pair of person and rating (swap the people and items).

Item-Based Filtering

6. recommendations.calculateSimilarItems(recommendations.critics) : builds the item similarity dataset, run frequently enough to keep the item similarities up to date.
7. recommendations.getRecommendedItems(recommendations.critics,itemsim,'Toby') : gets recommendation for 'Toby' based on 'recommendation.critics' as rating dictionary and 'itemsim' as item similarity dataset.

MovieLens Dataset

8. prefs=recommendations.loadMovieLens( ) : load the dataset into prefs dictonary object.
9. recommendations.getRecommendations(prefs,'87')[0:30] : returns top 30 results of user based recommendations for user with id '87'.