victorkohler

with graph.as_default():
    '''
    Loss function: 
    -SUM ln σ(xui - xuj) + λ(w1)**2 + λ(w2)**2 + λ(w3)**2 ...

    ln = the natural log
    σ(xuij) = the sigmoid function of xuij.
    λ = lambda regularization value.
    ||W||**2 = the squared L2 norm of our model parameters.

victorkohler

#-------------------------
# TENSORFLOW GRAPH
#-------------------------

# Set up our Tensorflow graph
graph = tf.Graph()

def init_variable(size, dim, name=None):
    '''

victorkohler

#-------------
# HYPERPARAMS
#-------------

epochs = 50
batches = 30
num_factors = 64 # Number of latent features

# Independent lambda regularization values 

victorkohler

import tensorflow as tf
import pandas as pd
import numpy as np
import scipy.sparse as sp

from tqdm import tqdm

#---------------------------
# LOAD AND PREPARE THE DATA
#---------------------------

victorkohler

import sys
import pandas as pd
import numpy as np
import scipy.sparse as sparse
from scipy.sparse.linalg import spsolve
import random

from sklearn.preprocessing import MinMaxScaler

import implicit

victorkohler

import sys
import pandas as pd
import numpy as np
import scipy.sparse as sparse
from scipy.sparse.linalg import spsolve
import random

from sklearn.preprocessing import MinMaxScaler

import implicit # The Cython library

victorkohler

def nonzeros(m, row):
    for index in xrange(m.indptr[row], m.indptr[row+1]):
        yield m.indices[index], m.data[index]
      
      
def implicit_als_cg(Cui, features=20, iterations=20, lambda_val=0.1):
    user_size, item_size = Cui.shape

    X = np.random.rand(user_size, features) * 0.01

victorkohler

#------------------------------
# FIND SIMILAR ITEMS
#------------------------------

# Let's find similar artists to Jay-Z. 
# Note that this ID might be different for you if you're using
# the full dataset or if you've sliced it somehow. 
item_id = 10277

victorkohler

# Let's say we want to recommend artists for user with ID 2023
user_id = 2023

#------------------------------
# GET ITEMS CONSUMED BY USER
#------------------------------

# Let's print out what the user has listened to
consumed_idx = data_sparse[user_id,:].nonzero()[1].astype(str)

victorkohler

user_vecs, item_vecs = implicit_als(data_sparse, iterations=20, features=20, alpha_val=40)