def analysis():
user_data = read_file.read_checkins_file("unknown_user")
user_dic = {}
for user_id, total_location in user_data.items():
user_dic[user_id] = []
for daily_location in total_location:
for single_location in daily_location:
user_dic[user_id].append(single_location["location"])
home_dic = home_recommendation.get_home_location()
result = {}
result["home"] = []
result["visitor"] = []
result["non_visitor"] = []
for user_id, user_location in user_dic.items():
for single_location in user_location:
for location_id, location_content in home_dic.items():
if user_id not in result["home"]:
if single_location == location_id:
result["home"].append(user_id)
for user_id, total_location in user_data.items():
if user_id not in result["home"]:
if len(total_location) > nonvisitor_threshold:
result["non_visitor"].append(user_id)
else:
result["visitor"].append(user_id)
#pprint.pprint(result)
return result
def unknownUser_knownLocation():
# 100 unknown users with 533736 known location
new_dic = {}
unknown_user_data = read_file.read_checkins_file("unknown_user")
unknown_user_list = list(unknown_user_data.keys())
loc_list = []
for keyname, valuename in unknown_user_data.items():
for dailyvalue in valuename:
for eachvalue in dailyvalue:
eachloc = eachvalue["location"]
loc_list.append(eachloc)
new_dic[keyname] = loc_list
pprint.pprint(new_dic)
known_loc = read_file.read_known_location()
unknownUser_knownLoc = (len(unknown_user_data), len(known_loc)
) #(2,3,4)变为3维
maxtrix_unknownUser_knownLoc = np.zeros(unknownUser_knownLoc)
#pprint.pprint(np.zeros(unknownUser_knownLoc))
for username, localist in new_dic.items():
userindex = unknown_user_list.index(username)
for location in localist:
try:
locationindex = known_loc.index(location)
except:
continue
maxtrix_unknownUser_knownLoc[userindex, locationindex] += 1
pprint.pprint(maxtrix_unknownUser_knownLoc)
a = np.array([1, 2, 3, 4, 5])
np.savetxt("b.txt", a)
np.savetxt("matrix.txt", maxtrix_unknownUser_knownLoc)
def knownUser_knownLocation():
# 3884 unknown users with 533736 known location
new_dic = {}
known_user_data = read_file.read_checkins_file("known_user")
known_user_list = list(known_user_data.keys())
pprint.pprint(len(known_user_list))
loc_list = []
for keyname, valuename in known_user_data.items():
for dailyvalue in valuename:
for eachvalue in dailyvalue:
eachloc = eachvalue["location"]
loc_list.append(eachloc)
new_dic[keyname] = loc_list
pprint.pprint(new_dic)
known_loc = read_file.read_known_location()
knownUser_knownLoc = (len(known_user_data), len(known_loc))
maxtrix_knownUser_knownLoc = np.zeros(knownUser_knownLoc)
pprint.pprint(np.zeros(maxtrix_knownUser_knownLoc))
for username, localist in new_dic.items():
userindex = known_user_list.index(username)
for location in localist:
try:
locationindex = known_loc.index(location)
except:
continue
maxtrix_knownUser_knownLoc[userindex, locationindex] += 1
def recommend():
unknown_user_checkins_dict = read_file.read_checkins_file('unknown_user')
location_info_dict = read_file.read_location_file()
average_location_dict = get_average_location(unknown_user_checkins_dict,
location_info_dict)
candidate_location_distance_dict = compute_distance(
average_location_dict, location_info_dict)
result_rank = rank_candidate_location(candidate_location_distance_dict)
return result_rank
def recommend():
user_data = read_file.read_checkins_file("unknown_user")
user_list = user_data.keys()
candidate_list = candidate_popularity()
result = {}
for user in user_list:
result[user] = []
for candidate in candidate_list:
result[user].append(candidate)
#pprint.pprint(result)
return result
def get_user_home(home_dict):
unknown_user_checkins_dict = read_file.read_checkins_file('unknown_user')
home_list = list(home_dict.keys())
user_home_location_dict = {}
for user_name, user_dict in unknown_user_checkins_dict.items():
home_location = have_home(user_dict, home_list)
if home_location:
user_home_location_dict[user_name] = home_dict[home_location]
# print('user_home_location_dict',len(user_home_location_dict))
return user_home_location_dict
def get_unknown_user():
user_data = read_file.read_checkins_file("unknown_user")
user_dic = {}
for user_id, user_total_location in user_data.items():
user_time_list = []
for user_daily_location in user_total_location:
for user_location in user_daily_location:
if user_location["location"] == "?":
user_time_list.append(user_location["time"])
user_dic[user_id] = user_time_list
#pprint.pprint(user_dic)
return user_dic
def total_popularity():
read_data = {}
read_data = read_file.read_checkins_file()
total_location = read_data.values()
popularity = {}
for daily_data in total_location:
for daily_location in daily_data:
for location_data in daily_location:
if popularity.get(location_data["location"]) != None:
popularity[location_data["location"]] +=1
else:
popularity[location_data["location"]] = 1
return popularity
def knownUser_unKnownLocation():
# 3884 known users with 100 unknown location
new_dic = {}
known_user_data = read_file.read_checkins_file("known_user")
known_user_list = list(known_user_data.keys())
loc_list = []
pprint.pprint("test")
for keyname, valuename in known_user_data.items():
for dailyvalue in valuename:
for eachvalue in dailyvalue:
eachloc = eachvalue["location"]
loc_list.append(eachloc)
new_dic[keyname] = loc_list
pprint.pprint("test2")
##pprint.pprint(new_dic)
pprint.pprint("test3")
known_loc = read_file.read_location_file()
pprint.pprint("test4")
user_dict = {}
for keyname, listvalue in new_dic.items():
user_dict[keyname] = {}
for locKey in known_loc.keys():
user_dict[keyname][locKey] = 0
for value in listvalue:
user_dict[keyname][value] += 1
pprint.pprint("loop2")
pprint.pprint(user_dict["9448"])
unknown_loc = read_file.read_candidate_file()
knownUser_unknownLoc = (len(known_user_data), len(unknown_loc)
) #(2,3,4)变为3维
maxtrix_knownUser_unknownLoc = np.zeros(knownUser_unknownLoc)
#pprint.pprint(np.zeros(knownUser_unknownLoc))
##pprint.pprint("build zero matrix")
'''
def get_person_location_type():
user_data = read_file.read_checkins_file("unknown_user")
location_data = read_file.read_location_file()
user_location_dic = {}
for user_id,user_location in user_data.items():
user_location_dic[user_id] = []
for daily_location in user_location:
for location in daily_location:
if location["location"] != "?":
location_type = location_data[location["location"]]["type"]
if "\x1a\x1a" in location_type:
location_type = "Café"
user_location_dic[user_id].append(location_type)
for user_id,user_location_type in user_location_dic.items():
user_location_dic[user_id] = {}
for single_location_type in user_location_type:
user_location_dic[user_id][single_location_type] = user_location_type.count(single_location_type)
#pprint.pprint(user_location_dic)
return user_location_dic
def recommend():
# 3884 known users with 100 unknown location
new_dic = {}
unknown_dic = {}
known_user_data = read_file.read_checkins_file("known_user")
# pprint.pprint("test")
unknown_user_data = read_file.read_checkins_file("unknown_user")
for keyname, valuename in known_user_data.items():
loc_list = []
for dailyvalue in valuename:
for eachvalue in dailyvalue:
eachloc = eachvalue["location"]
loc_list.append(eachloc)
new_dic[keyname] = loc_list
#pprint.pprint(new_dic)
for keyname, valuename in unknown_user_data.items():
loc_list = []
for dailyvalue in valuename:
for eachvalue in dailyvalue:
eachloc = eachvalue["location"]
loc_list.append(eachloc)
unknown_dic[keyname] = loc_list
knn_rank_dic = {}
for cur_userid, cur_loclist in unknown_dic.items():
rank_user_sim = {}
for next_userid, next_loclist in new_dic.items():
count = 0
#if next_userid != "1006" and next_userid != "1437":
for eachLoc in cur_loclist:
if eachLoc in next_loclist:
count += 1
rank_user_sim[next_userid] = count
#pprint.pprint(len(rank_user_sim))
rank_list = []
for rankUser, rankValue in rank_user_sim.items():
rank_list.append([rankValue, rankUser])
rank_list.sort(reverse=True)
knn_list = []
for time in range(0, 5):
knn_list.append(rank_list[time][1])
knn_rank_dic[cur_userid] = knn_list
#pprint.pprint(knn_rank_dic)
candidate_data = read_file.read_candidate_file()
candidate_dic = {}
for user_id, user_candidate in knn_rank_dic.items():
candidate_dic[user_id] = {}
for candidate in candidate_data:
count = 0
for user_data in user_candidate:
for user_location in new_dic[user_data]:
if candidate == user_location:
count += 1
candidate_dic[user_id][candidate] = count
result = {}
for user_id, user_candidate in candidate_dic.items():
candidate_rank_list = []
result_list = []
for candidate_id, candidate_value in user_candidate.items():
candidate_rank_list.append([candidate_value, candidate_id])
candidate_rank_list.sort(reverse=True)
for i in range(0, len(candidate_rank_list)):
result_list.append(candidate_rank_list[i][1])
result[user_id] = result_list
# pprint.pprint(result["95132"])
return result
