""" Excercise 3 """ import recommendations similar_critics = recommendations.calculate_similar_items(recommendations.movies) def get_precomputed_recommendedations(prefs, similar, user): similar_prefs = {} for other in similar: similar_prefs[other] = prefs[other] similar_prefs[user] = prefs[user] return recommendations.get_recommendations(similar_prefs, user) # The precompute recommendations tend to match up extremely well with the more # laboriously computed recommendations. Probably because they are generated the # same way. Larger datasets would probably make this match less exact, but # would give huge speed improvements, since all algorithms tend to be fast for # small n.
def main(): prefs = loadMovieLens() itemsim = recommendations.calculate_similar_items(prefs, n=50) print (itemsim) print(recommendations.get_recommended_items(prefs, itemsim, '87')[0:30])
""" Excercise 3 """ import recommendations similar_critics = recommendations.calculate_similar_items( recommendations.movies) def get_precomputed_recommendedations(prefs, similar, user): similar_prefs = {} for other in similar: similar_prefs[other] = prefs[other] similar_prefs[user] = prefs[user] return recommendations.get_recommendations(similar_prefs, user) # The precompute recommendations tend to match up extremely well with the more # laboriously computed recommendations. Probably because they are generated the # same way. Larger datasets would probably make this match less exact, but # would give huge speed improvements, since all algorithms tend to be fast for # small n.
# -*- coding: utf-8 -*- import recommendations reload(recommendations) info = recommendations.calculate_similar_items(recommendations.critics) # print info print recommendations.get_recommended_items(recommendations.critics, info, 'Jack Matthews')
# -*- coding: utf-8 -*- import recommendations reload(recommendations) # from deliciousrec import * # delusers = initialize_user_dict('programing') # fill_items(delusers) # # print delusers data = recommendations.load_movie_lens() print '==================== 사용자 ====================' print data['87'] print '==================== 추천 ====================' print recommendations.getRecommendations(data, '87')[0:30] print '==================== 항목 기반 ====================' item_sim = recommendations.calculate_similar_items(data, rank=50) print recommendations.get_recommended_items(data, item_sim, '87')[0:30]
# Lib imports import recommendations, time # Function's execution t1 = time.time() prefs=recommendations.load_movie_lens() t2 = time.time() print "Recomendations: {}".format(recommendations.get_recommendations(prefs,'87')[0:30]) t3 = time.time() itemsim=recommendations.calculate_similar_items(prefs,50) t4 = time.time() print "Recommended items: {}".format(recommendations.get_recommended_items(prefs,itemsim,'87')[0:30]) t5 = time.time() print "\nExecution times" print "-----------------" print "Load dataset: {} seconds".format(t2-t1) print "User based filtering: {} seconds".format(t3-t2) print "Calculate similar item: {} seconds".format(t4-t3) print "Item based filtering: {} seconds".format(t5-t4)