CN103116696B - Personnel based on the mobile phone location data of sparse sampling reside place recognition methods - Google Patents

Abstract

The personnel that the present invention relates to a kind of mobile phone location data based on sparse sampling reside place recognition methods, it is characterized in that, step is: step 1, by target cities gridding; Step 2, set up the mapping relations of mobile phone location data and grid; The probability of occurrence of step 3, counting user and the frequency of occurrences; Step 4, respectively space-time cluster is carried out to each cellphone subscriber's frequency of occurrences and probability of occurrence, occur to be incorporated in Time and place the grid that feature is similar; The resident place of step 5, each cellphone subscriber identifies.Advantage of the present invention is: leverage fully on existing mobile communications network resource, with the mobile phone location data of magnanimity for input, adopt virtual grid technology, realize setting up mapping relations fast, reduce the operand with map match, simultaneously based on the frequency of occurrences and probability, adopt space-time cluster, realize the recognition mechanism that user resides place.This invention can be used for automatically identifying personnel's permanent residence dot information.

Description

Sparse sampling-based personnel resident location identification method for mobile phone positioning data

Technical Field

The invention relates to a sparse sampling-based personnel resident location identification method for mobile phone positioning data, which is used for mining valuable personnel travel tracks from massive mobile phone positioning data and deducing information of personnel resident locations, can serve traffic planning, traffic demand management, traffic policy evaluation and traffic operation management, and belongs to the technical field of traffic planning and management methods.

Background

With the rapid economic growth, the rapid infrastructure construction and the frequent change of land utilization in recent years, the trip characteristics and the trip laws of people in almost all cities in China are changed frequently. Particularly, with the popularization of the high-traffic, high-speed and high-reliability travel modes such as rail transit, the travel in a large spatial range of a large number of people in a short time can be completely met, and if the travel is concentrated on the peak in the morning and at the evening, great pressure is inevitably caused on traffic. Therefore, accurate and reliable personnel resident places in the urban area range are one of basic data for providing decision support quantitative analysis for traffic operation management measures, policies and the like, and the automatic acquisition of traffic basic data and travel characteristic data by an informatization means is particularly urgent.

With the advancement of modern communication technologies and the increasing level of services, wireless communication networks are capable of covering all areas of space that are accessible to humans. Meanwhile, with the popularization of mobile phone terminals, the ownership rate and the utilization rate of the mobile phone reach a quite high proportion. Valuable personnel travel characteristic information can be mined from mass mobile phone positioning data, resident places can be identified, and the method and the system can serve for traffic planning, traffic demand management, traffic policy evaluation and traffic operation management.

Disclosure of Invention

The invention aims to provide a method for acquiring the resident place condition of a mobile phone user.

In order to achieve the above object, the technical solution of the present invention is to provide a method for identifying a resident location of a person based on sparsely sampled mobile phone positioning data, which is characterized by comprising the steps of:

step 1, dividing a target city into a columns and b rows of grids by using grids with the size of N meters by N meters, and producing grid information of each grid, wherein the grid information at least comprises grid numbers, X-axis coordinates of grid center points and Y-axis coordinates of the grid center points;

step 2, mobile phone positioning data of all mobile phone users in a target city, which are sparsely sampled in a certain time period, are obtained, each piece of mobile phone positioning data has X-axis coordinate information and Y-axis coordinate information, a mutual mapping relation between each piece of mobile phone positioning data and each grid is established, the total number r of grids which appear in each mobile phone user is obtained, and for the h-th mobile phone positioning data of the ith mobile phone user, the h-th mobile phone positioning data of the ith mobile phone user is compared with the grid information of each grid, if the h-th mobile phone positioning data meets the following requirements: x is more than or equal to gx-N/2 and less than gx + N/2, Y is more than or equal to gy-N/2 and less than gy + N/2, wherein gx and gy are respectively an X-axis coordinate of a grid central point and a Y-axis coordinate of the grid central point, X and Y are respectively X-axis coordinate information and Y-axis coordinate information carried by the h-th mobile phone positioning data, and then the h-th mobile phone positioning data and the grid where the gx and the gy meet the conditions are in mapping relation;

step 3, taking n days in a certain time period as analysis days, dividing each analysis day into m analysis periods, counting the occurrence frequency of the same analysis period of each analysis day of each mobile phone user in each grid according to the mapping relation between the mobile phone positioning data and the grids, and counting the occurrence frequency of the tth analysis period of all analysis days of the ith mobile phone user in the jth gridWherein:

counting the occurrence probability of the same analysis period in each grid of all analysis days of each mobile phone user, wherein the occurrence probability of the tth analysis period in the jth grid of all analysis days of the ith mobile phone user $P_i^{\mathrm{jt}}=\frac{N_i^{\mathrm{jt}}}{\underset{j=1}{\overset{r}{\mathrm{\Σ}}}{M}_i^j},$ $M_i^j=\underset{t=1}{\overset{m}{\mathrm{\Σ}}}{N}_i^{\mathrm{jt}};$

Counting the occurrence probability of each mobile phone user in each grid on all analysis days, and counting the occurrence probability of the ith mobile phone user in the jth grid on all analysis days

Step 4, performing space-time clustering on the occurrence frequency and the occurrence probability of each mobile phone user respectively to combine grids with similar occurrence characteristics in time and space, and for the ith mobile phone user, the method comprises the following steps:

step 4.1, time similarity analysis:

defining a grid j₁And grid j₂Time similarity judgment parameter in the t-th analysis period $T_i^{j_1{j}_2}=\mathrm{\Σ}{H}_i^{{\mathrm{tj}}_1{j}_2},$ $H_i^{{\mathrm{tj}}_1{j}_2}=\left\{\begin{array}{cc}1& F_i^{{\mathrm{tj}}_1}=F_i^{{\mathrm{tj}}_2}\\ 0& F_i^{{\mathrm{tj}}_1}\≠F_i^{{\mathrm{tj}}_2}\end{array}\right.,$ $F_i^{{\mathrm{tj}}_1}=\left\{\begin{array}{cc}1& Q_i^{j_1}\≥{\mathrm{\θ}}_1\mathrm{and}{P}_i^{j_{1}t}\≥{\mathrm{\θ}}_2\\ 0& \mathrm{else}\end{array}\right.,$ $F_i^{{\mathrm{tj}}_2}=\left\{\begin{array}{cc}1& Q_i^{j_2}\≥{\mathrm{\θ}}_1\mathrm{and}{P}_i^{j_{2}t}\≥{\mathrm{\θ}}_2\\ 0& \mathrm{else}\end{array}\right.,$ Wherein,andrespectively for the ith mobile phone user on the grid j on all the analysis days₁And grid j₂The probability of occurrence of (a) is,andrespectively for the ith mobile phone user in the t analysis period in the grid j₁And grid j₂Is in the case ofThen the ith mobile phone user is considered to be in the grid j₁And grid j₂Has a time appearance characteristic of time similarity of theta₁、θ₂And theta₃Is a preset experience threshold;

step 4.2, spatial similarity analysis:

defining a grid j₁And grid j₂Judgment parameters with similar space rangesWherein (x)₁，y₁) And (x)₂，y₂) Are respectively a grid j₁And grid j₂Coordinates of the center point of (1), ifθ₄If the experience threshold is preset, the ith mobile phone user is considered to be in the grid j₁And grid j₂Have spatial similarity over a spatial range of (a);

step 4.3, merging grids with time similarity and space similarity in grids appeared by the ith mobile phone user into a grid set to obtain a plurality of grid sets, wherein each grid set is used as an analysis gridAnd (3) respectively reserving grids which cannot be combined and other grids as an analysis grid, calculating the occurrence probability of each analysis grid in all analysis days as the analysis occurrence probability, and if the current analysis grid is the jth grid which cannot be combined, determining the analysis occurrence probability Z of the current analysis grid_iFor the occurrence probability of the ith mobile phone user in the grid on all analysis days, if the current analysis grid is a grid set consisting of y grids, the analysis occurrence probability of the current analysis grid $Z_i=\frac{\underset{x=1}{\overset{y}{\mathrm{\Σ}}}{M}_i^{j_x}}{\underset{j=1}{\overset{r}{\mathrm{\Σ}}}{M}_i^j};$

Step 5, identifying the resident place of each mobile phone user, and for the ith mobile phone user, the steps are as follows:

step 5.1, defining resident location identification parameters and resident location identification parameters of the ith mobile phone userG (j) is the identification parameter of the jth grid, $G\left(j\right)=\left\{\begin{array}{cc}1& {Z_i}^j>{\mathrm{\θ}}_5\\ 0& {Z_i}^j\≤{\mathrm{\θ}}_5\end{array}\right.,$ Z_i ^jfor the analysis occurrence probability of the analysis grid corresponding to the jth grid, θ₅Is a preset experience threshold; calculating the total sampling rate of each mobile phone user and the total sampling rate of the ith mobile phone user

And 5.2, identifying the resident place of the mobile phone user, and for the ith mobile phone user:

if (S)_i＜θ₆) Or (S)_i≥θ₆And C_i0), the resident place cannot be judged;

if S_i≥θ₆And C_iIf the probability of the analysis of the ith mobile phone user is more than or equal to 1, the front C with the maximum probability of the analysis of the ith mobile phone user is determined_iTaking an analysis grid as a resident place, if the analysis grid is a grid set, taking any grid in the grid set as the resident place, and theta₆Is a preset empirical threshold.

Preferably, in the step 5.2, if the analysis grid is a grid set, the ith mobile phone user with the highest probability of appearing in all the analysis days in the grid set is taken as a resident place.

The invention has the advantages that: the method is characterized in that the existing mobile communication network resources are fully relied on, massive mobile phone positioning data are used as input, a virtual grid technology is adopted, the mapping relation is quickly established, the calculation amount matched with a map is reduced, and meanwhile, the recognition mechanism of the resident place of a user is realized by adopting space-time clustering based on the occurrence frequency and probability. The invention can be used for automatically identifying the resident place information of the personnel.

Drawings

FIG. 1 is a schematic diagram of a target urban grid;

FIG. 2A is a histogram of the occurrence probability for the case where stationary points cannot be determined;

FIG. 2B is a histogram of the occurrence probability of merging grids in the case where the stationary points cannot be determined;

FIG. 3A is a histogram of the probability of occurrence in the average mode;

FIG. 3B is a histogram of the probability of occurrence after merging the grids in the average mode;

FIG. 4A is a histogram of probability of occurrence for unimodal mode;

FIG. 4B is a histogram of the probability of occurrence of merging grids in unimodal mode;

FIG. 5A is a histogram of the probability of occurrence in the bimodal mode;

FIG. 5B is a histogram of the probability of occurrence after merging the grids in bimodal mode;

FIG. 6A is a histogram of the probability of occurrence in multimodal mode;

FIG. 6B is a histogram of the probability of occurrence after merging the grids for a multimodal mode;

FIG. 7 is a flow chart of the present invention.

Detailed Description

In order to make the invention more comprehensible, preferred embodiments are described in detail below with reference to the accompanying drawings.

With reference to fig. 7, the present invention provides a sparse sampling-based method for identifying a resident location of a person using mobile phone positioning data, which comprises the steps of:

step 1, dividing the target city into a columns and b rows of grids by using the grids with the size of 500 m × 500 m, for example, producing grid information of each grid, wherein the format of the grid information is shown in table 1, and the grid information includes grid number gridd, X-axis coordinate gx of a grid center point and Y-axis coordinate gy of the grid center point, and in this embodiment, specific values of the grid information are shown in table 2.

Field(s)	Type (B)	Description of the invention
			gridid	number	Grid numbering
gx	number	Grid center point x coordinate
			gy	number	Grid center point y coordinate

TABLE 1

gridid	gx	gy
			1	120.87	30.61
2	620.87	30.61
			...	...	...

TABLE 2

Step 2, mobile phone positioning data of all mobile phone users in a target city, which are sparsely sampled in a certain time period, are obtained, each piece of mobile phone positioning data has X-axis coordinate information and Y-axis coordinate information, a mutual mapping relation between each piece of mobile phone positioning data and each grid is established, the total number r of grids which appear in each mobile phone user is obtained, and for the h-th mobile phone positioning data of the ith mobile phone user, the h-th mobile phone positioning data of the ith mobile phone user is compared with the grid information of each grid, if the h-th mobile phone positioning data meets the following requirements: gx-500/2 is more than or equal to X and less than gx +500/2, and gy-500/2 is more than or equal to Y and less than gy +500/2, wherein gx and gy are respectively an X-axis coordinate of a grid center point and a Y-axis coordinate of the grid center point, X and Y are respectively X-axis coordinate information and Y-axis coordinate information carried by the h-th mobile phone positioning data, and then the h-th mobile phone positioning data is mapped with the grid where gx and gy meet the conditions, and in this embodiment, the data shown in table 3 is generated;

TABLE 3

Step 3, taking n days in a certain time period as analysis days, dividing each analysis day into m analysis periods, counting the occurrence frequency of the same analysis period of each analysis day of each mobile phone user in each grid according to the mapping relation between the mobile phone positioning data and the grids, and counting the occurrence frequency of the tth analysis period of all analysis days of the ith mobile phone user in the jth gridWherein:

counting the occurrence probability of the same analysis period in each grid of all analysis days of each mobile phone user, wherein the occurrence probability of the tth analysis period in the jth grid of all analysis days of the ith mobile phone user In this embodiment, data as shown in table 4 can be obtained:

mobile phone number	Grid mesh	Time period	N	P
					1	1	1	51	3％
1	1	2	60	4％
					1	1	...	...	...
1	1	12	0	0
					1	1	13	0	0
1	1	...	...	...
					1	1	t	66	4％
...	...	...	...	...

TABLE 4

Counting the occurrence probability of each mobile phone user in each grid on all analysis days, wherein the ith mobile phone user

Probability of occurrence in jth grid for all days analyzedIn this example, data as shown in Table 5 were obtained:

mobile phone number	Grid mesh	M	Q
				1	1	700	47％
1	2	5	0.3％
				1	...	...	...
1	j	1	0.1％
				...	...	...	...

TABLE 5

Step 4, performing space-time clustering on the occurrence frequency and the occurrence probability of each mobile phone user respectively to combine grids with similar occurrence characteristics in time and space, and for the ith mobile phone user, the method comprises the following steps:

step 4.1, time similarity analysis:

defining a grid j₁And grid j₂Time similarity judgment parameter in the t-th analysis period $T_i^{j_1{j}_2}=\mathrm{\Σ}{H}_i^{{\mathrm{tj}}_1{j}_2},$ $H_i^{{\mathrm{tj}}_1{j}_2}=\left\{\begin{array}{cc}1& F_i^{{\mathrm{tj}}_1}=F_i^{{\mathrm{tj}}_2}\\ 0& F_i^{{\mathrm{tj}}_1}\≠F_i^{{\mathrm{tj}}_2}\end{array}\right.,$ $F_i^{{\mathrm{tj}}_1}=\left\{\begin{array}{cc}1& Q_i^{j_1}\≥{\mathrm{\θ}}_1\mathrm{and}{P}_i^{j_{1}t}\≥{\mathrm{\θ}}_2\\ 0& \mathrm{else}\end{array}\right.,$ $F_i^{{\mathrm{tj}}_2}=\left\{\begin{array}{cc}1& Q_i^{j_2}\≥{\mathrm{\θ}}_1\mathrm{and}{P}_i^{j_{2}t}\≥{\mathrm{\θ}}_2\\ 0& \mathrm{else}\end{array}\right.,$ Wherein,andrespectively for the ith mobile phone user on the grid j on all the analysis days₁And grid j₂The probability of occurrence of (a) is,andrespectively for the ith mobile phone user in the t analysis period in the grid j₁And grid j₂Is in the case ofThen the ith mobile phone user is considered to be in the grid j₁And grid j₂Has a time appearance characteristic of time similarity of theta₁、θ₂And theta₃In this embodiment, the values of the preset empirical threshold are shown in table 6:

parameter(s)	Threshold value
		θ1	0.2
θ2	0.015
		θ3	10

Step 4.2, spatial similarity analysis:

defining a grid j₁And grid j₂Judgment parameters with similar space rangesWherein (x)₁，y₁) And (x)₂，y₂) Are respectively a grid j₁And grid j₂Coordinates of the center point of (1), ifθ₄Is a predetermined empirical threshold, theta in this embodiment₄If the length is 1000 m, the ith mobile phone user is considered to be in the grid j₁And grid j₂Have spatial similarity over a spatial range of (a);

step 4.3, merging grids which have time similarity and space similarity simultaneously in grids appeared by the ith mobile phone user into a grid set to obtain a plurality of grid sets, wherein each grid set is used as an analysis grid, and other grids which cannot be merged are respectively used as analysis gridsOne analysis grid is reserved, the occurrence probability of each analysis grid in all analysis days is calculated as the analysis occurrence probability, and if the current analysis grid is the jth grid which can not be combined, the analysis occurrence probability Z of the current analysis grid is_iFor the occurrence probability of the ith mobile phone user in the grid on all analysis days, if the current analysis grid is a grid set consisting of y grids, the analysis occurrence probability of the current analysis grid $Z_i=\frac{\underset{x=1}{\overset{y}{\mathrm{\Σ}}}{M}_i^{j_x}}{\underset{j=1}{\overset{r}{\mathrm{\Σ}}}{M}_i^j};$

Step 5, identifying the resident place of each mobile phone user, and for the ith mobile phone user, the steps are as follows:

step 5.1, defining resident location identification parameters and resident location identification parameters of the ith mobile phone userG (j) is the identification parameter of the jth grid, $G\left(j\right)=\left\{\begin{array}{cc}1& {Z_i}^j>{\mathrm{\θ}}_5\\ 0& {Z_i}^j\≤{\mathrm{\θ}}_5\end{array}\right.,$ Z_i ^jfor the analysis occurrence probability of the analysis grid corresponding to the jth grid, θ₅The preset empirical threshold is taken to be 0.2 in this embodiment; calculating the total sampling rate of each mobile phone user and the total sampling rate of the ith mobile phone user $S_i=\frac{\underset{j=1}{\overset{r}{\mathrm{\Σ}}}{M}_i^j}{m\×n};$

And 5.2, identifying the resident place of the mobile phone user, wherein for the ith mobile phone user, theta₆For the preset empirical threshold, in this example, it is 0.05:

if S_i＜θ₆If yes, the ith mobile phone user behavior is not active enough and noneWhen the method determines the location where the location is resident, the histogram of the occurrence probability and the histogram after grid merging are shown in fig. 2A and fig. 2B, respectively.

If S_i≥θ₆And C_iIf 0, the ith mobile phone user belongs to the average mode, the travel range is wide, the travel is frequent, the resident place cannot be determined, and the histogram of the occurrence probability and the histogram of the grid combination are shown in fig. 3A and fig. 3B, respectively.

If S_i≥θ₆And C_iIf the analysis grid is a grid set, the grid set takes the analysis grid with the highest occurrence probability of the ith mobile phone user in all the analysis days as the resident location, and the histogram of the occurrence probability and the histogram after grid combination are respectively shown in fig. 4A and 4B.

If S_i≥θ₆And C_iIf the analysis grid is a grid set, the grid set takes the first two analysis grids with the highest occurrence probability of the ith mobile phone user in all the analysis days as the resident places, and the histogram of the occurrence probability and the histogram after grid combination are respectively shown in fig. 5A and 5B.

If S_i≥θ₆And C_iIf the probability of the ith mobile phone user is more than or equal to 3, the ith mobile phone user belongs to a multimodal mode, and the front C with the maximum analysis occurrence probability of the ith mobile phone user is used_iEach analysis grid is taken as a resident place, if the analysis grid is a grid set, the grid set takes the place with the highest occurrence probability in all the analysis days of the ith mobile phone user as the resident place, and the histogram of the occurrence probability and the histogram after grid combination are respectively shown in fig. 6A and fig. 6B.

Claims (2)

1. A personnel resident place identification method based on mobile phone positioning data of sparse sampling is characterized by comprising the following steps:

step 1, dividing a target city into a columns and b rows of grids by using grids with the size of N meters by N meters, and producing grid information of each grid, wherein the grid information at least comprises grid numbers, X-axis coordinates of grid center points and Y-axis coordinates of the grid center points;

step 2, mobile phone positioning data of all mobile phone users in a target city, which are sparsely sampled in a certain time period, are obtained, each piece of mobile phone positioning data has X-axis coordinate information and Y-axis coordinate information, a mutual mapping relation between each piece of mobile phone positioning data and each grid is established, the total number r of grids which appear in each mobile phone user is obtained, and for the h-th mobile phone positioning data of the ith mobile phone user, the h-th mobile phone positioning data of the ith mobile phone user is compared with the grid information of each grid, if the h-th mobile phone positioning data meets the following requirements: x is more than or equal to gx-N/2 and less than gx + N/2, Y is more than or equal to gy-N/2 and less than gy + N/2, wherein gx and gy are respectively an X-axis coordinate of a grid central point and a Y-axis coordinate of the grid central point, X and Y are respectively X-axis coordinate information and Y-axis coordinate information carried by the h-th mobile phone positioning data, and then the h-th mobile phone positioning data and the grid where the gx and the gy meet the conditions are in mapping relation;

step 3, taking n days in a certain time period as analysis days, dividing each analysis day into m analysis periods, counting the occurrence frequency of the same analysis period of each analysis day of each mobile phone user in each grid according to the mapping relation between the mobile phone positioning data and the grids, and counting the occurrence frequency of the tth analysis period of all analysis days of the ith mobile phone user in the jth gridWherein:

counting the occurrence probability of the same analysis period in each grid of all analysis days of each mobile phone user, wherein the occurrence probability of the tth analysis period in the jth grid of all analysis days of the ith mobile phone user $P_i^{jt}=\frac{N_i^{jt}}{\underset{j=1}{\overset{r}{\Σ}}{M}_i^j},$ Wherein, $M_i^j=\underset{t=1}{\overset{m}{\Σ}}{N}_i^{jt};$

counting the occurrence probability of each mobile phone user in each grid on all analysis days, and counting the occurrence probability of the ith mobile phone user in the jth grid on all analysis days

Step 4, performing space-time clustering on the occurrence frequency and the occurrence probability of each mobile phone user respectively to combine grids with similar occurrence characteristics in time and space, and for the ith mobile phone user, the method comprises the following steps:

step 4.1, time similarity analysis:

defining a grid j₁And grid j₂Time similarity judgment parameter in the t-th analysis period ${T_i}^{j_1{j}_2}={{\mathrm{\ΣH}}_i}^{{\mathrm{tj}}_1{j}_2},$ ${H_i}^{{\mathrm{tj}}_1{j}_2}=\left\{\begin{array}{cc}1& {F_i}^{{\mathrm{tj}}_1}={F_i}^{{\mathrm{tj}}_2}\\ 0& {F_i}^{{\mathrm{tj}}_1}\≠{F_i}^{{\mathrm{tj}}_2}\end{array}\right.,$ ${F_i}^{{\mathrm{tj}}_1}=\left\{\begin{array}{cc}1& \begin{array}{ccc}{Q_i}^{j_1}\≥{\mathrm{\θ}}_1& and& {P_i}^{j_{1}t}\≥{\mathrm{\θ}}_2\end{array}\\ 0& else\end{array}\right.,$ ${F_i}^{{\mathrm{tj}}_2}=\left\{\begin{array}{cc}1& \begin{array}{ccc}{Q_i}^{j_2}\≥{\mathrm{\θ}}_1& and& {P_i}^{j_{2}t}\≥{\mathrm{\θ}}_2\end{array}\\ 0& else\end{array}\right.,$

Wherein,andrespectively for the ith mobile phone user on the grid j on all the analysis days₁And grid j₂Probability of occurrence inAndrespectively for the ith mobile phone user in the t analysis period in the grid j₁And grid j₂Is in the case ofThen the ith mobile phone user is considered to be in the grid j₁And grid j₂Has a time appearance characteristic of time similarity of theta₁、θ₂And theta₃Is a preset experience threshold;

step 4.2, spatial similarity analysis:

defining a grid j₁And grid j₂Judgment parameters with similar space rangesWherein (x)₁，y₁) And (x)₂，y₂) Are respectively a grid j₁And grid j₂Coordinates of the center point of (1), ifθ₄If the experience threshold is preset, the ith mobile phone user is considered to be in the grid j₁And grid j₂Have spatial similarity over a spatial range of (a);

step 4.3, merging grids which have time similarity and space similarity simultaneously in grids appeared by the ith mobile phone user into a grid set to obtain a plurality of grid sets, taking each grid set as an analysis grid, respectively reserving other grids which cannot be merged as an analysis grid, calculating the appearance probability of each analysis grid in all analysis days as the analysis appearance probability, and if the current analysis grid is the jth grid which cannot be merged, then the analysis appearance probability Z of the current analysis grid_iFor the occurrence probability of the ith mobile phone user in the grid on all analysis days, if the current analysis grid is a grid set consisting of y grids, the analysis occurrence probability of the current analysis grid $Z_i=\frac{\underset{x=1}{\overset{y}{\mathrm{\Σ}}}{M}_i^{j_x}}{\underset{j=1}{\overset{r}{\mathrm{\Σ}}}{M}_i^j};$

Step 5, identifying the resident place of each mobile phone user, and for the ith mobile phone user, the steps are as follows:

step 5.1, defining resident location identification parameters and resident location identification parameters of the ith mobile phone userFor the identification parameter of the jth mesh, $G\left(j\right)=\left\{\begin{array}{cc}1& {Z_i}^j>{\mathrm{\θ}}_5\\ 0& {Z_i}^j\≤{\mathrm{\θ}}_5\end{array}\right.,$ Z_i ^jfor the analysis occurrence probability of the analysis grid corresponding to the jth grid, θ₅Is a preset experience threshold; calculating the total sampling rate of each mobile phone user and the total sampling rate of the ith mobile phone user

And 5.2, identifying the resident place of the mobile phone user, and for the ith mobile phone user:

if (S)_i＜θ₆) Or (S)_i≥θ₆And C_i0), the resident place cannot be judged;

if S_i≥θ₆And C_iIf the probability of the analysis of the ith mobile phone user is more than or equal to 1, the front C with the maximum probability of the analysis of the ith mobile phone user is determined_iTaking an analysis grid as a resident place, if the analysis grid is a grid set, taking any grid in the grid set as the resident place, and theta₆Is a preset empirical threshold.

2. The method as claimed in claim 1, wherein in step 5.2, if the analysis grid is a grid set, the resident location is selected from the grid set that has the highest probability of occurrence in all analysis days of the ith mobile phone user.