workshops_recommender_systems/.ipynb_checkpoints/P4. Matrix Factorization-checkpoint.ipynb

Self made SVD

import helpers
import pandas as pd
import numpy as np
import scipy.sparse as sparse
from collections import defaultdict
from itertools import chain
import random

train_read=pd.read_csv('./Datasets/ml-100k/train.csv', sep='\t', header=None)
test_read=pd.read_csv('./Datasets/ml-100k/test.csv', sep='\t', header=None)
train_ui, test_ui, user_code_id, user_id_code, item_code_id, item_id_code = helpers.data_to_csr(train_read, test_read)

# Done similarly to https://github.com/albertauyeung/matrix-factorization-in-python
from tqdm import tqdm

class SVD():
    
    def __init__(self, train_ui, learning_rate, regularization, nb_factors, iterations):
        self.train_ui=train_ui
        self.uir=list(zip(*[train_ui.nonzero()[0],train_ui.nonzero()[1], train_ui.data]))
        
        self.learning_rate=learning_rate
        self.regularization=regularization
        self.iterations=iterations
        self.nb_users, self.nb_items=train_ui.shape
        self.nb_ratings=train_ui.nnz
        self.nb_factors=nb_factors
        
        self.Pu=np.random.normal(loc=0, scale=1./self.nb_factors, size=(self.nb_users, self.nb_factors))
        self.Qi=np.random.normal(loc=0, scale=1./self.nb_factors, size=(self.nb_items, self.nb_factors))

    def train(self, test_ui=None):
        if test_ui!=None:
            self.test_uir=list(zip(*[test_ui.nonzero()[0],test_ui.nonzero()[1], test_ui.data]))
            
        self.learning_process=[]
        pbar = tqdm(range(self.iterations))
        for i in pbar:
            pbar.set_description(f'Epoch {i} RMSE: {self.learning_process[-1][1] if i>0 else 0}. Training epoch {i+1}...')
            np.random.shuffle(self.uir)
            self.sgd(self.uir)
            if test_ui==None:
                self.learning_process.append([i+1, self.RMSE_total(self.uir)])
            else:
                self.learning_process.append([i+1, self.RMSE_total(self.uir), self.RMSE_total(self.test_uir)])
    
    def sgd(self, uir):
        
        for u, i, score in uir:
            # Computer prediction and error
            prediction = self.get_rating(u,i)
            e = (score - prediction)
            
            # Update user and item latent feature matrices
            Pu_update=self.learning_rate * (e * self.Qi[i] - self.regularization * self.Pu[u])
            Qi_update=self.learning_rate * (e * self.Pu[u] - self.regularization * self.Qi[i])
            
            self.Pu[u] += Pu_update
            self.Qi[i] += Qi_update
        
    def get_rating(self, u, i):
        prediction = self.Pu[u].dot(self.Qi[i].T)
        return prediction
    
    def RMSE_total(self, uir):
        RMSE=0
        for u,i, score in uir:
            prediction = self.get_rating(u,i)
            RMSE+=(score - prediction)**2
        return np.sqrt(RMSE/len(uir))
    
    def estimations(self):
        self.estimations=\
        np.dot(self.Pu,self.Qi.T)

    def recommend(self, user_code_id, item_code_id, topK=10):
        
        top_k = defaultdict(list)
        for nb_user, user in enumerate(self.estimations):
            
            user_rated=self.train_ui.indices[self.train_ui.indptr[nb_user]:self.train_ui.indptr[nb_user+1]]
            for item, score in enumerate(user):
                if item not in user_rated and not np.isnan(score):
                    top_k[user_code_id[nb_user]].append((item_code_id[item], score))
        result=[]
        # Let's choose k best items in the format: (user, item1, score1, item2, score2, ...)
        for uid, item_scores in top_k.items():
            item_scores.sort(key=lambda x: x[1], reverse=True)
            result.append([uid]+list(chain(*item_scores[:topK])))
        return result
    
    def estimate(self, user_code_id, item_code_id, test_ui):
        result=[]
        for user, item in zip(*test_ui.nonzero()):
            result.append([user_code_id[user], item_code_id[item], 
                           self.estimations[user,item] if not np.isnan(self.estimations[user,item]) else 1])
        return result

model=SVD(train_ui, learning_rate=0.005, regularization=0.02, nb_factors=100, iterations=40)
model.train(test_ui)

Epoch 39 RMSE: 0.7481223595239049. Training epoch 40...: 100%|██████████| 40/40 [02:09<00:00,  3.25s/it]

import matplotlib.pyplot as plt

df=pd.DataFrame(model.learning_process).iloc[:,:2]
df.columns=['epoch', 'train_RMSE']
plt.plot('epoch', 'train_RMSE', data=df, color='blue')
plt.legend()

<matplotlib.legend.Legend at 0x7f40f450bf98>

import matplotlib.pyplot as plt

df=pd.DataFrame(model.learning_process[10:], columns=['epoch', 'train_RMSE', 'test_RMSE'])
plt.plot('epoch', 'train_RMSE', data=df, color='blue')
plt.plot('epoch', 'test_RMSE', data=df, color='yellow', linestyle='dashed')
plt.legend()

<matplotlib.legend.Legend at 0x7f40f148f710>

Saving and evaluating recommendations

model.estimations()

top_n=pd.DataFrame(model.recommend(user_code_id, item_code_id, topK=10))

top_n.to_csv('Recommendations generated/ml-100k/Self_SVD_reco.csv', index=False, header=False)

estimations=pd.DataFrame(model.estimate(user_code_id, item_code_id, test_ui))
estimations.to_csv('Recommendations generated/ml-100k/Self_SVD_estimations.csv', index=False, header=False)

import evaluation_measures as ev

estimations_df=pd.read_csv('Recommendations generated/ml-100k/Self_SVD_estimations.csv', header=None)
reco=np.loadtxt('Recommendations generated/ml-100k/Self_SVD_reco.csv', delimiter=',')

ev.evaluate(test=pd.read_csv('./Datasets/ml-100k/test.csv', sep='\t', header=None),
            estimations_df=estimations_df, 
            reco=reco,
            super_reactions=[4,5])

943it [00:00, 4506.01it/s]

	RMSE	MAE	precision	recall	F_1	F_05	precision_super	recall_super	NDCG	mAP	MRR	LAUC	HR	Reco in test	Test coverage	Shannon	Gini
0	0.915304	0.719016	0.100848	0.042228	0.051191	0.067885	0.092275	0.07073	0.104366	0.049606	0.192999	0.517831	0.465536	0.867869	0.150072	3.847796	0.972676

import evaluation_measures as ev

dir_path="Recommendations generated/ml-100k/"
super_reactions=[4,5]
test=pd.read_csv('./Datasets/ml-100k/test.csv', sep='\t', header=None)

ev.evaluate_all(test, dir_path, super_reactions)

943it [00:00, 4962.65it/s]
943it [00:00, 4020.02it/s]
943it [00:00, 3974.62it/s]
943it [00:00, 4763.58it/s]
943it [00:00, 4203.99it/s]
943it [00:00, 4230.96it/s]
943it [00:00, 3909.05it/s]
943it [00:00, 4215.47it/s]
943it [00:00, 4293.30it/s]
943it [00:00, 4389.83it/s]
943it [00:00, 4410.76it/s]

	Model	RMSE	MAE	precision	recall	F_1	F_05	precision_super	recall_super	NDCG	mAP	MRR	LAUC	HR	Reco in test	Test coverage	Shannon	Gini
0	Ready_LightFMpureMF	7.953192	7.462008	0.334464	0.219997	0.217225	0.254981	0.233798	0.266952	0.398778	0.263058	0.629129	0.607709	0.913043	1.000000	0.275613	5.085818	0.913665
0	Ready_LightFM	162.707436	160.855483	0.340827	0.217682	0.217990	0.258010	0.243884	0.260663	0.403850	0.268266	0.637590	0.606568	0.898197	1.000000	0.351371	5.366291	0.885046
0	Self_P3	3.702446	3.527273	0.282185	0.192092	0.186749	0.216980	0.204185	0.240096	0.339114	0.204905	0.572157	0.593544	0.875928	1.000000	0.077201	3.875892	0.974947
0	Ready_ImplicitALS	3.266101	3.065824	0.255037	0.188653	0.176852	0.201189	0.166631	0.214925	0.305908	0.172546	0.523871	0.591709	0.889714	1.000000	0.502886	5.722957	0.827507
0	Self_TopPop	2.508258	2.217909	0.188865	0.116919	0.118732	0.141584	0.130472	0.137473	0.214651	0.111707	0.400939	0.555546	0.765642	1.000000	0.038961	3.159079	0.987317
0	Ready_LightFMcontent	182.471340	180.405210	0.160339	0.101224	0.102198	0.121074	0.102682	0.112455	0.180079	0.087429	0.337825	0.547572	0.704136	0.974973	0.264791	4.909893	0.926201
0	Self_SVD	0.915304	0.719016	0.100848	0.042228	0.051191	0.067885	0.092275	0.070730	0.104366	0.049606	0.192999	0.517831	0.465536	0.867869	0.150072	3.847796	0.972676
0	Ready_Baseline	0.949459	0.752487	0.091410	0.037652	0.046030	0.061286	0.079614	0.056463	0.095957	0.043178	0.198193	0.515501	0.437964	1.000000	0.033911	2.836513	0.991139
0	Self_GlobalAvg	1.125760	0.943534	0.061188	0.025968	0.031383	0.041343	0.040558	0.032107	0.067695	0.027470	0.171187	0.509546	0.384942	1.000000	0.025974	2.711772	0.992003
0	Ready_Random	1.514355	1.216383	0.049735	0.022300	0.025782	0.033598	0.028219	0.021751	0.054383	0.021119	0.133978	0.507680	0.339343	0.986957	0.177489	5.088670	0.907676
0	Self_BaselineUI	0.967585	0.762740	0.000954	0.000170	0.000278	0.000463	0.000644	0.000189	0.000752	0.000168	0.001677	0.496424	0.009544	0.600530	0.005051	1.803126	0.996380

Embeddings

x=np.array([[1,2],[3,4]])
display(x)
x/np.linalg.norm(x, axis=1)[:,None]

array([[1, 2],
       [3, 4]])

array([[0.4472136 , 0.89442719],
       [0.6       , 0.8       ]])

item=random.choice(list(set(train_ui.indices)))

embeddings_norm=model.Qi/np.linalg.norm(model.Qi, axis=1)[:,None] # we do not mean-center here
# omitting normalization also makes sense, but items with a greater magnitude will be recommended more often

similarity_scores=np.dot(embeddings_norm,embeddings_norm[item].T)
top_similar_items=pd.DataFrame(enumerate(similarity_scores), columns=['code', 'score'])\
.sort_values(by=['score'], ascending=[False])[:10]

top_similar_items['item_id']=top_similar_items['code'].apply(lambda x: item_code_id[x])

items=pd.read_csv('./Datasets/ml-100k/movies.csv')

result=pd.merge(top_similar_items, items, left_on='item_id', right_on='id')

result

	code	score	item_id	id	title	genres
0	44	1.000000	45	45	Eat Drink Man Woman (1994)	Comedy, Drama
1	855	0.966812	856	856	Night on Earth (1991)	Comedy, Drama
2	1403	0.966571	1404	1404	Withnail and I (1987)	Comedy
3	112	0.966115	113	113	Horseman on the Roof, The (Hussard sur le toit...	Drama
4	955	0.965365	956	956	Nobody's Fool (1994)	Drama
5	1222	0.965232	1223	1223	King of the Hill (1993)	Drama
6	60	0.964481	61	61	Three Colors: White (1994)	Drama
7	535	0.963322	536	536	Ponette (1996)	Drama
8	1102	0.962597	1103	1103	Trust (1990)	Comedy, Drama
9	713	0.962459	714	714	Carrington (1995)	Drama, Romance

project task 5: implement SVD on top baseline (as it is in Surprise library)

# making changes to our implementation by considering additional parameters in the gradient descent procedure 
# seems to be the fastest option
# please save the output in 'Recommendations generated/ml-100k/Self_SVDBaseline_reco.csv' and
# 'Recommendations generated/ml-100k/Self_SVDBaseline_estimations.csv'

Ready-made SVD - Surprise implementation

SVD

import helpers
import surprise as sp
import imp
imp.reload(helpers)

algo = sp.SVD(biased=False) # to use unbiased version

helpers.ready_made(algo, reco_path='Recommendations generated/ml-100k/Ready_SVD_reco.csv',
          estimations_path='Recommendations generated/ml-100k/Ready_SVD_estimations.csv')

Generating predictions...
Generating top N recommendations...
Generating predictions...

SVD biased - on top baseline

import helpers
import surprise as sp
import imp
imp.reload(helpers)

algo = sp.SVD() # default is biased=True

helpers.ready_made(algo, reco_path='Recommendations generated/ml-100k/Ready_SVDBiased_reco.csv',
          estimations_path='Recommendations generated/ml-100k/Ready_SVDBiased_estimations.csv')

Generating predictions...
Generating top N recommendations...
Generating predictions...

import imp
imp.reload(ev)

import evaluation_measures as ev
dir_path="Recommendations generated/ml-100k/"
super_reactions=[4,5]
test=pd.read_csv('./Datasets/ml-100k/test.csv', sep='\t', header=None)

ev.evaluate_all(test, dir_path, super_reactions)

943it [00:00, 4110.82it/s]
943it [00:00, 4014.43it/s]
943it [00:00, 3946.85it/s]
943it [00:00, 4832.12it/s]
943it [00:00, 4090.40it/s]
943it [00:00, 4152.16it/s]
943it [00:00, 4456.22it/s]
943it [00:00, 3943.66it/s]
943it [00:00, 4298.65it/s]
943it [00:00, 4243.05it/s]
943it [00:00, 4391.40it/s]
943it [00:00, 4528.23it/s]
943it [00:00, 4419.43it/s]

	Model	RMSE	MAE	precision	recall	F_1	F_05	precision_super	recall_super	NDCG	mAP	MRR	LAUC	HR	Reco in test	Test coverage	Shannon	Gini
0	Ready_LightFMpureMF	7.953192	7.462008	0.334464	0.219997	0.217225	0.254981	0.233798	0.266952	0.398778	0.263058	0.629129	0.607709	0.913043	1.000000	0.275613	5.085818	0.913665
0	Ready_LightFM	162.707436	160.855483	0.340827	0.217682	0.217990	0.258010	0.243884	0.260663	0.403850	0.268266	0.637590	0.606568	0.898197	1.000000	0.351371	5.366291	0.885046
0	Self_P3	3.702446	3.527273	0.282185	0.192092	0.186749	0.216980	0.204185	0.240096	0.339114	0.204905	0.572157	0.593544	0.875928	1.000000	0.077201	3.875892	0.974947
0	Ready_ImplicitALS	3.266101	3.065824	0.255037	0.188653	0.176852	0.201189	0.166631	0.214925	0.305908	0.172546	0.523871	0.591709	0.889714	1.000000	0.502886	5.722957	0.827507
0	Self_TopPop	2.508258	2.217909	0.188865	0.116919	0.118732	0.141584	0.130472	0.137473	0.214651	0.111707	0.400939	0.555546	0.765642	1.000000	0.038961	3.159079	0.987317
0	Ready_LightFMcontent	182.471340	180.405210	0.160339	0.101224	0.102198	0.121074	0.102682	0.112455	0.180079	0.087429	0.337825	0.547572	0.704136	0.974973	0.264791	4.909893	0.926201
0	Ready_SVD	0.951475	0.750225	0.099470	0.051407	0.056004	0.070229	0.088197	0.083166	0.115422	0.053515	0.253329	0.522434	0.522800	0.996713	0.216450	4.424505	0.952962
0	Self_SVD	0.915304	0.719016	0.100848	0.042228	0.051191	0.067885	0.092275	0.070730	0.104366	0.049606	0.192999	0.517831	0.465536	0.867869	0.150072	3.847796	0.972676
0	Ready_Baseline	0.949459	0.752487	0.091410	0.037652	0.046030	0.061286	0.079614	0.056463	0.095957	0.043178	0.198193	0.515501	0.437964	1.000000	0.033911	2.836513	0.991139
0	Ready_SVDBiased	0.937841	0.739906	0.079427	0.032570	0.039804	0.053022	0.071030	0.050639	0.088490	0.039308	0.201565	0.512929	0.425239	0.997031	0.170996	4.167051	0.963929
0	Self_GlobalAvg	1.125760	0.943534	0.061188	0.025968	0.031383	0.041343	0.040558	0.032107	0.067695	0.027470	0.171187	0.509546	0.384942	1.000000	0.025974	2.711772	0.992003
0	Ready_Random	1.514355	1.216383	0.049735	0.022300	0.025782	0.033598	0.028219	0.021751	0.054383	0.021119	0.133978	0.507680	0.339343	0.986957	0.177489	5.088670	0.907676
0	Self_BaselineUI	0.967585	0.762740	0.000954	0.000170	0.000278	0.000463	0.000644	0.000189	0.000752	0.000168	0.001677	0.496424	0.009544	0.600530	0.005051	1.803126	0.996380