pd03.py

import numpy as np
import pandas as pd
from sklearn import preprocessing
from sklearn.metrics.pairwise import cosine_similarity
from scipy.spatial import distance as d

le = preprocessing.LabelEncoder()

user_profiles = pd.read_csv("user_profile.csv", delimiter='\t')
target_users = pd.read_csv("target_users.csv", delimiter='\t')
item_profiles = pd.read_csv("item_profile.csv", delimiter='\t')
item_raccomandabili = pd.read_csv('item_racc.csv', delimiter='\t').drop('Unnamed: 0',1)

users_r = pd.merge(user_profiles,target_users,on='user_id')
items_r = pd.merge(item_profiles,item_raccomandabili,on='id')

common_attributes = np.intersect1d(item_profiles.columns.values,user_profiles.columns.values).tolist()

users = users_r[common_attributes]
items = items_r[common_attributes]

users = users.fillna(0)
items = items.fillna(0)

users['country'] = users['country'].replace(0,'null')
users['country'] = le.fit_transform(users['country'])

items['country'] = items['country'].replace(0,'null')
items['country'] = le.fit_transform(items['country'])

items = items.astype(int)
users = users.astype(int)

def encode_feature(dataframe, to_encode):
    hot = preprocessing.OneHotEncoder(dtype='int')

    for feature in to_encode:
        a = hot.fit_transform(dataframe[feature].values.reshape((-1,1))).toarray()
        for i in range(np.shape(a)[1]):
            dataframe[feature + str(i)] = a[:,i]
    return

encode_feature(items,common_attributes)
encode_feature(users,common_attributes)
items = items.drop(common_attributes,1)
users = users.drop(common_attributes,1)

for attribute in users.columns.difference(items.columns).tolist():
    items[attribute] = pd.DataFrame(np.zeros((2486,)))

items = items.astype(int).sort_index(axis=1)
users = users.astype(int).sort_index(axis=1)

# cm = np.intersect1d(items.columns.values,users.columns.values).tolist()
# items = items[cm]
# users = users[cm]

def similarity(u,v):
    SHRINK = 7
    a = np.dot(u,v)
    b = np.dot(np.norm(u),np.norm(v)) + SHRINK
    return a/b

ratings = pd.DataFrame(columns=('user_id', 'item_id', 'rating'))

j=0
simil_matrix = np.ndarray(shape=(users.shape[0],items.shape[0]))
for u in range(users.shape[0]):
    user_id = users_r.ix[[u]]['user_id'].values[0]
    user = users.ix[[u]].values
    print(str(u))
    for i in range(items.shape[0]):
        item_id = items_r.ix[[i]]['id'].values[0]
        item = items.ix[[i]].values
        simil_matrix[u][i] = 1 - d.cosine(user,item)