def Show_Users_on_10_FMT():
indx = 0
friend_data = get_friends_data()
for ky in friend_data:
print ky
indx += 1
if indx > 20:
break
print "..."
def Show_Users_on_10_FMT():
indx = 0
friend_data = get_friends_data()
for ky in friend_data:
print ky
indx += 1
if indx > 20:
break
print "..."
def Generate_Simulating_Data_on_10_FMT(user_id):
filelocation = "./FMT10/"
friend_data = get_friends_data()
raw_data = get_user_item_matrix()
generate_simulating_data(raw_data, friend_data, user_id, 30, 10,
filelocation)
def JPH_on_10_FMT():
f_s = [10, 20, 30, 40, 50]
friend_data = get_friends_data()
raw_data = get_user_item_matrix()
jhk_friends(raw_data, friend_data, f_s)
def Single_Stranger_Influence_on_10_FMT():
friend_data = get_friends_data()
raw_data = get_user_item_matrix()
filename = "stranger_dif_10fmt_"
single_influence(raw_data, friend_data, False, filename)
def Single_Friend_Influence_on_10_FMT():
friend_data = get_friends_data()
raw_data = get_user_item_matrix()
filename = "friend_dif_10fmt_"
single_influence(raw_data, friend_data, True, filename)
def MAE_on_10_FMT():
friend_data = get_friends_data()
raw_data = get_user_item_matrix()
f_ts = [(10, 10), (20, 10), (30, 10), (40, 10), (50, 10)]
calculate_cosine_friends_strangers(raw_data, friend_data, f_ts)
def MAE_on_10_FMT():
friend_data = get_friends_data()
raw_data = get_user_item_matrix()
f_ts = [(10,10),(20,10),(30,10),(40,10),(50,10)]
calculate_cosine_friends_strangers(raw_data, friend_data, f_ts)
color='k',
marker='>',
alpha=0.8,
label="friends")
plt.title('Cosine Similarity')
plt.axis('tight')
plt.ylabel("Similarity")
plt.xlabel("User ID")
plt.legend(loc='lower right')
plt.show()
if __name__ == '__main__':
raw_data = get_user_item_matrix()
friend_data = get_friends_data()
sim_mat = get_similarity_matrix(raw_data, friend_data)
f_sim = get_friends_similarity(sim_mat, friend_data)
t_sim = get_strangers_similarity(sim_mat, friend_data)
idx = 1
f_sims = {}
t_sims = {}
for ky in f_sim:
f_sims[idx] = f_sim[ky]
t_sims[idx] = t_sim[ky]
idx += 1
draw_similarity(f_sims, t_sims)
def create_model():
raw_data = get_user_item_matrix()
friend_data = get_friends_data()
i_model = FriendsModel(raw_data, friend_data)
return i_model
def get_args():
return
def show_dict(the_dict):
total_v = 0.0
for ky in the_dict.keys():
total_v += the_dict[ky]
print len(the_dict.keys())
print total_v / len(the_dict.keys())
if __name__ == '__main__':
#raw_data = get_original_UI(1)#get_original_user_item_matrix() #get_user_item_matrix()
friend_data_1 = get_friends_data()
raw_data_1 = get_user_item_matrix()
#calculate_jph(raw_data_1,friend_data_1)
calculate_cosine_friends_strangers(raw_data_1, friend_data_1)
#reputation_data = get_reputation()
#calculate_reputation_rmse_friends(raw_data,friend_data,reputation_data)
#calculate_reputation_rmse_strangers(raw_data,friend_data,reputation_data)
#calculate_nmf_rmse_global(raw_data)
#var_arr = calculate_rate_average_global(raw_data,friend_data)
#total = 0.0
#for u_id in var_arr.keys():
# var1 = var_arr[u_id]
# total += var1
#avg = total/float(len(var_arr.keys()))
def Generate_Simulating_Data_on_10_FMT(user_id):
filelocation = "./FMT10/"
friend_data = get_friends_data()
raw_data = get_user_item_matrix()
generate_simulating_data(raw_data,friend_data,user_id,30,10,filelocation)
def JPH_on_10_FMT():
f_s = [10, 20 ,30, 40, 50]
friend_data = get_friends_data()
raw_data = get_user_item_matrix()
jhk_friends(raw_data, friend_data, f_s)
def Single_Stranger_Influence_on_10_FMT():
friend_data = get_friends_data()
raw_data = get_user_item_matrix()
filename = "stranger_dif_10fmt_"
single_influence(raw_data,friend_data,False,filename)
def Single_Friend_Influence_on_10_FMT():
friend_data = get_friends_data()
raw_data = get_user_item_matrix()
filename = "friend_dif_10fmt_"
single_influence(raw_data,friend_data,True,filename)
def get_args():
return
def show_dict(the_dict):
total_v = 0.0
for ky in the_dict.keys():
total_v+=the_dict[ky]
print len(the_dict.keys())
print total_v/len(the_dict.keys())
if __name__ == '__main__':
#raw_data = get_original_UI(1)#get_original_user_item_matrix() #get_user_item_matrix()
friend_data_1 = get_friends_data()
raw_data_1 = get_user_item_matrix()
#calculate_jph(raw_data_1,friend_data_1)
calculate_cosine_friends_strangers(raw_data_1,friend_data_1)
#reputation_data = get_reputation()
#calculate_reputation_rmse_friends(raw_data,friend_data,reputation_data)
#calculate_reputation_rmse_strangers(raw_data,friend_data,reputation_data)
#calculate_nmf_rmse_global(raw_data)
#var_arr = calculate_rate_average_global(raw_data,friend_data)
#total = 0.0
#for u_id in var_arr.keys():
# var1 = var_arr[u_id]
# total += var1
#avg = total/float(len(var_arr.keys()))
plt.axis('tight')
plt.show()
def show_statistic(ivector):
sumx = {}
for item in ivector:
if item in sumx:
sumx[item] += 1
else:
sumx[item] = 1
colors = list("rgbcmyk")
plt.scatter(sumx.keys(),sumx.values(),color=colors.pop())
plt.legend(sumx.keys())
plt.show()
def run_single_influence(raw_data, friend_data):
f_n = 10
t_n = 11
ratio = 0.8
esvlidation = EsoricsSingleUserValidation(5,raw_data,friend_data,f_n,t_n,ratio)
results = esvlidation.cross_validate()
return results
if __name__ == '__main__':
friend_data = get_friends_data()
raw_data = get_user_item_matrix()
results = run_single_influence(raw_data,friend_data)
show_data(results)
def Pure_Friend_Influence_on_10_FMT():
friend_data = get_friends_data()
raw_data = get_user_item_matrix()
filename = "pure_friend_dif_10fmt_"
pure_single_friend_influence(raw_data, friend_data, filename)
def create_model():
raw_data = get_user_item_matrix()
friend_data = get_friends_data()
i_model = FriendsModel(raw_data, friend_data)
return i_model
def Pure_Friend_Influence_on_10_FMT():
friend_data = get_friends_data()
raw_data = get_user_item_matrix()
filename = "pure_friend_dif_10fmt_"
pure_single_friend_influence(raw_data,friend_data,filename)
