#!/usr/bin/env python
# coding: UTF-8
from __future__ import division
from poirank.transition import init_data, trans
import pylab
poi_axis, axis_poi, data = init_data(tuple(range(2, 30)))
#tensor = trans(data, len(axis_poi))
# print "data: ", data
data_length = len(data)
print "data_length: ", data_length
def poi_coorelation(order):
esum = 0.0
for i in range(data_length):
esum = esum + data[i]
expectation = esum / data_length
print "expectation: ", expectation
vsum = 0.0
for j in range(data_length):
vsum += pow((data[j] - expectation), 2)
print "vsum: ", vsum
variance = vsum / data_length
tsum = 0.0
for k in range(data_length - order):
tsum += (data[k] - expectation) * (data[k + order] - expectation)
print "tsum: ", tsum
ar = tsum / vsum
#!/usr/bin/env python
# coding: UTF-8
from __future__ import division
from poirank.transition import init_data
import pylab
import numpy as np
'''
贝叶斯信息准则
在实际中不太好操作，对于状态数很大的马尔科夫链，极大似然值小而惩罚值过大，容易过拟合，从而阶数始终为1。
'''
poi_axis, axis_poi, data = init_data((6,))
#poi_axis, axis_poi, data = init_data(tuple(range(0, 182)))
data_length = len(data)
state_num = len(axis_poi)
print "key: ", axis_poi.keys()
print "state_num: ", state_num
# tensor = []
# tensor[1] = [[0 for i in range(state_num)] for j in range(state_num)]
# tensor[2] = [[[0 for i in range(state_num)] for j in range(state_num)] for k in range(state_num)]
# tensor[3] = [[[[0 for i in range(state_num)] for j in range(state_num)] for k in range(state_num)] for l in range(state_num)]
# tensor[4] = [[[[[0 for i in range(state_num)] for j in range(state_num)] for k in range(state_num)] for l in range(state_num)] for m in range(state_num)]
def bic_1(data):
tensor1 = [0 for i in range(state_num)]
tensor2 = [[0 for l in range(state_num)] for m in range(state_num)]
for index in range(0, data_length-1):
check_list = data[index:index+1]
check_list_plus_one = data[index:index+2]
# print check_list, check_list_plus_one
#!/usr/bin/env python
# coding: UTF-8
from __future__ import division
import math
import numpy as np
from poirank.transition import init_data, trans
import pylab
from numpy import *
# 以紫荆园为中心
z_latitude = 40.010428
z_longtitude = 116.322341
poi_axis, axis_poi, data = init_data(tuple(range(0, 182)), return_gps=True)
data_length = len(data)
print "data_length: ", data_length
#print data
# 计算两个经纬度之间的距离，单位千米
def calculate_distance(lat1, lng1, lat2, lng2):
earth_radius = 6378.137
rad_lat1 = rad(lat1)
rad_lat2 = rad(lat2)
a = rad_lat1 - rad_lat2
b = rad(lng1) - rad(lng2)
s = 2 * math.asin(
math.sqrt(math.pow(math.sin(a / 2), 2) + math.cos(rad_lat1) * math.cos(rad_lat2) * math.pow(math.sin(b / 2), 2)))
s *= earth_radius
if s < 0:
meta_data = (index, x[index])
sort_data.append(meta_data)
sort_data.sort(key=lambda x: x[1], reverse=True)
res = []
for i in range(0, len(sort_data)):
res.append(
(i + 1, axis_poi[axis_poi2[sort_data[i][0]]], sort_data[i][1]))
return res
if __name__ == '__main__':
i_distance = 2
current_position = (116.361684, 39.951409) # poi为1231
poi_axis, axis_poi, data, name2type = init_data(
tuple(range(0,
182)), return_gps=True, return_poi_type=True) # (0, 182)
print "当前位置: ", axis_poi[1231]
print "name2type: ", name2type
count_list = []
count_set = set()
for i in range(0, len(data)):
distance = calculate_distance(data[i][2], data[i][1],
current_position[1], current_position[0])
if distance <= i_distance:
count_list.append(data[i])
count_set.add(data[i][0])
print "count_list: ", len(count_list)
print count_set
sort_data = []
for index in range(0, len(x)):
meta_data = (index, x[index])
sort_data.append(meta_data)
sort_data.sort(key=lambda x: x[1], reverse=True)
res = []
for i in range(0, len(sort_data)):
res.append((i+1, axis_poi[axis_poi2[sort_data[i][0]]], sort_data[i][1]))
return res
if __name__ == '__main__':
i_distance = 2
current_position = (116.361684, 39.951409) # poi为1231
poi_axis, axis_poi, data, name2type = init_data(tuple(range(0, 182)), return_gps=True, return_poi_type=True) # (0, 182)
print "当前位置: ", axis_poi[1231]
print "name2type: ", name2type
count_list = []
count_set = set()
for i in range(0, len(data)):
distance = calculate_distance(data[i][2], data[i][1], current_position[1], current_position[0])
if distance <= i_distance:
count_list.append(data[i])
count_set.add(data[i][0])
print "count_list: ", len(count_list)
print count_set
print len(count_set)
dimensionality = len(count_set)
#!/usr/bin/env python
# coding: UTF-8
from __future__ import division
from poirank.transition import init_data
import pylab
import numpy as np
'''
贝叶斯信息准则
在实际中不太好操作，对于状态数很大的马尔科夫链，极大似然值小而惩罚值过大，容易过拟合，从而阶数始终为1。
'''
poi_axis, axis_poi, data = init_data((6, ))
#poi_axis, axis_poi, data = init_data(tuple(range(0, 182)))
data_length = len(data)
state_num = len(axis_poi)
print "key: ", axis_poi.keys()
print "state_num: ", state_num
# tensor = []
# tensor[1] = [[0 for i in range(state_num)] for j in range(state_num)]
# tensor[2] = [[[0 for i in range(state_num)] for j in range(state_num)] for k in range(state_num)]
# tensor[3] = [[[[0 for i in range(state_num)] for j in range(state_num)] for k in range(state_num)] for l in range(state_num)]
# tensor[4] = [[[[[0 for i in range(state_num)] for j in range(state_num)] for k in range(state_num)] for l in range(state_num)] for m in range(state_num)]
def bic_1(data):
tensor1 = [0 for i in range(state_num)]
tensor2 = [[0 for l in range(state_num)] for m in range(state_num)]
for index in range(0, data_length - 1):
check_list = data[index:index + 1]
check_list_plus_one = data[index:index + 2]
# print check_list, check_list_plus_one
#!/usr/bin/env python
# coding: UTF-8
from __future__ import division
from poirank.transition import init_data, trans
import pylab
poi_axis, axis_poi, data = init_data(tuple(range(2,30)))
#tensor = trans(data, len(axis_poi))
# print "data: ", data
data_length = len(data)
print "data_length: ", data_length
def poi_coorelation(order):
esum = 0.0
for i in range(data_length):
esum = esum + data[i]
expectation = esum / data_length
print "expectation: ", expectation
vsum = 0.0
for j in range(data_length):
vsum += pow((data[j] - expectation), 2)
print "vsum: ", vsum
variance = vsum / data_length
tsum = 0.0
for k in range(data_length - order):
tsum += (data[k] - expectation) * (data[k + order] - expectation)
print "tsum: ", tsum
ar = tsum / vsum
x_values = []
while (block_width < 3):
block_num = int(30 // block_width)
block_matrix = [[
i + 1 for i in range(j * block_num, (j + 1) * block_num)
] for j in range(0, block_num)]
block_transition = [[0 for i in range(0, block_num * block_num)]
for j in range(0, block_num * block_num)]
print block_matrix
print block_transition
block_array_list = {}
for i in range(1, block_num * block_num + 1):
block_array_list[i] = []
poi_axis, axis_poi, data = init_data(tuple(range(0, 182)),
return_gps=True) # (0, 182)
max_latitude = 0
min_latitude = data[0][2]
max_longtitude = 0
min_longtitude = data[0][1]
for i in range(len(data)):
if data[i][2] > max_latitude:
max_latitude = data[i][2]
if data[i][2] < min_latitude:
min_latitude = data[i][2]
if data[i][1] > max_longtitude:
max_longtitude = data[i][1]
if data[i][1] < min_longtitude:
min_longtitude = data[i][1]
