CN109508360B - Geographical multivariate stream data space-time autocorrelation analysis method based on cellular automaton - Google Patents

Geographical multivariate stream data space-time autocorrelation analysis method based on cellular automaton Download PDF

Info

Publication number
CN109508360B
CN109508360B CN201811318953.7A CN201811318953A CN109508360B CN 109508360 B CN109508360 B CN 109508360B CN 201811318953 A CN201811318953 A CN 201811318953A CN 109508360 B CN109508360 B CN 109508360B
Authority
CN
China
Prior art keywords
time
data
cellular
space
geographical
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201811318953.7A
Other languages
Chinese (zh)
Other versions
CN109508360A (en
Inventor
陈江平
熊志鹏
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Wuhan University WHU
Original Assignee
Wuhan University WHU
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Wuhan University WHU filed Critical Wuhan University WHU
Priority to CN201811318953.7A priority Critical patent/CN109508360B/en
Priority to AU2018101946A priority patent/AU2018101946A4/en
Publication of CN109508360A publication Critical patent/CN109508360A/en
Application granted granted Critical
Publication of CN109508360B publication Critical patent/CN109508360B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/12—Classification; Matching
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/182—Network patterns, e.g. roads or rivers
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/10—Terrestrial scenes
    • G06V20/188—Vegetation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/048—Activation functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10032—Satellite or aerial image; Remote sensing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20084—Artificial neural networks [ANN]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30181—Earth observation
    • G06T2207/30188—Vegetation; Agriculture

Abstract

The invention provides a space-time autocorrelation analysis method of geographical multivariate stream data based on a cellular automaton, which adopts an improved cellular automaton dynamic model to express the space-time property and the complexity of the geographical data, considers the cellular (geographical region) conversion rule and the spatial heterogeneity of asynchronous evolution, and can more accurately analyze the nonlinear structure geographical multivariate stream data based on a complex network. The method analyzes the cellular unit, extracts various influence factors, can more accurately obtain the cellular automaton model parameters, and is accurate and high in efficiency; the conversion rule obtained by the ANN algorithm has dynamic property compared with the fixed conversion rule of the whole model, and can describe and accord with the actual conversion condition of the cells; the correlation among the cells is expressed according to Moran's I, the space-time distribution condition of the geographic data is better and more clearly reflected, and therefore the subsequent space-time data model simulation and prediction are more conveniently carried out, and the simulation and prediction precision is higher.

Description

Geographical multivariate stream data space-time autocorrelation analysis method based on cellular automaton
Technical Field
The invention belongs to a geospatial statistical analysis method, and particularly relates to the fields of cellular automata, geographical multivariate stream data processing, geospatial correlation analysis and the like.
Background
With the continuous development and popularization of geographic information technology, remote sensing technology, internet of things and mobile terminals in recent years, the multivariate geographic data with spatial reference form massive multivariate geographic stream data after decades of rapid evolution. These stream data have the characteristics of sequential, large-scale, rapid and continuous arrival, and can be gathered into a dynamic data set which can grow infinitely along with the time, and also can be called space-time stream data, such as traffic data, temperature real-time monitoring data and the like.
At present, the spatial autocorrelation analysis technology can better reveal the intrinsic spatial characteristics of a research object, and is widely applied to analysis and monitoring of spatial distribution characteristics of environments and disasters. However, it can be found from the literature studied by various national scholars that the model of the geographical multivariate stream data is mainly constructed based on a time series model, only the time autocorrelation is considered, and the geospatial heterogeneity and the correlation among multivariate variables are not considered. As in the related literature, the wave group limited invented "a method for analyzing data using distributed 3D dynamic space" patent application no: CN201410529971.5 publication no: in CN104298743A, in a certain time period, collecting spatial data of a dynamic body, storing data information at a certain time point at a specified time interval, and being used for behavior analysis and behavior judgment of the dynamic body; the invention relates to a method for analyzing the correlation between social events and passenger traffic demands in the invention patent of Beijing university of transportation, which has the following patent application numbers: CN201711403609.3 publication no: in CN108133291A, a fusion model is constructed on the designed attributes for different types of events by using a machine learning algorithm, and the variation of traffic demand caused by events is predicted. The spatial autocorrelation analysis method is mainly applied to discrete space-time data, the space-time data is divided into space data and time sequence data, correlation analysis is carried out on the two kinds of data independently, and interaction between time and space is ignored.
Therefore, the invention innovatively provides a solution for the spatiotemporal autocorrelation analysis of the geographical multivariate stream data based on the cellular automata, adopts an improved cellular automata dynamic model to express the spatiotemporal property and the complexity of the geographical data, considers the cellular (geographical region) conversion rule and the spatial heterogeneity of asynchronous evolution, and can more accurately analyze the nonlinear structure geographical multivariate stream data based on the complex network.
Disclosure of Invention
The invention aims to realize the analysis of the spatiotemporal autocorrelation of the geographical multivariate stream data. Aiming at the characteristics that data has time dimension and continuous arrival, a geographical multivariate stream data expression model based on a cellular automaton is constructed, and a spatiotemporal autocorrelation analysis method for improving the Moran's I index is provided. The method can reveal the space-time evolution characteristics of the geographical multivariate flow data, provides a solution considering the continuous expression of time and space dimensions for the prediction analysis in the aspects of real-time temperature monitoring, traffic monitoring and the like, and has important significance for processing geographical data.
The invention mainly comprises the following steps:
step 1, classifying original geographical multivariate stream data to obtain raster data containing various categories;
step 2, improving a cellular automaton model, and establishing the cellular automaton model which accords with the geographic multi-source flow data characteristics;
step 3, defining each grid in the grid data as a cell unit, and extracting an influence factor in each cell unit;
step 4, determining the state condition of the cells and the neighborhood relationship among the cells;
step 5, randomly extracting attribute data in a plurality of cells as training sample data;
step 6, repeatedly training the training sample data by using an ANN algorithm in machine learning by considering the diversity and the space-time property of the data until an optimal solution is obtained, and taking the obtained output result as a parameter required by the cellular automaton model;
step 7, taking the extracted influence factors as an input layer, adopting the result obtained by training as an initial weight value, and mining a conversion rule in each cell by using an ANN algorithm to obtain an initial conversion probability of converting the cell type into another type;
step 8, extracting the maximum value by using the output layer data obtained by training, namely the conversion probability data, determining whether the cell type changes at the next moment, and determining the trend of the type of conversion;
step 9, calculating a space-time weight value between the cells by adopting a molar neighborhood mode of the cell model;
step 10, improving a space-time Moran's I analysis variable space-time structure according to the continuous characteristics of the geographical multivariate stream data and the model expression forms on three levels of space position, attribute and time;
and 11, considering the correlation among the cells, combining the space-time weight matrix calculated in the step 9 and the Moran's I index improved in the step 10 with a cellular automaton model established based on geographical multivariate stream data to perform simulation analysis and prediction to form a combined model with a space-time autocorrelation analysis function on the basis of the cellular automaton model.
Furthermore, the cellular automata model conforming to the characteristics of the geographic multisource flow data in the step 2 is defined as follows,
a ═ { F, L, G, S, N, R } equation (2)
In the above formula, F ═ { F } represents attribute information data that can be expressed in a quantized manner; l ═ { L } represents the distribution of geographic phenomena; g is the cellular partitioning of the geospatial space; n ═ c1,c2,...,cnIs a cellular neighbor; s ═ ScRepresents a state variable of the cell; r ═ { R } represents a set of rules, otherwise known as a set of evolution functions.
Further, in step 5, the maximum value and the minimum value are normalized, so that the obtained training sample data are all between 0 and 1.
Further, it is determined whether the cell type is changed at the next time in step 8, and the calculation formula of the trend of determining which type is converted is as follows,
pt(i,j)＝(1+(-lnγ)a)×pg×con(St(i,j))×Ωt(i, j) formula (5)
In the formula, pt(i, j) is the transition probability of the class represented by a certain cell; 1+ (-ln γ)aRandom items are expressed, so that the simulation result is more consistent with the actual situation; p is a radical ofgIs the global transition probability; con (S)t(i, j))) represents a constraint condition for a cell unit; omegat(i, j) represents a neighborhood function, representing the effect of the neighborhood on the cell transition probability.
Further, there are two expression ways for the modified Moran's I index in step 10, the specific expression is as follows,
moran's I index OSTI in the form of a global spatio-temporal streamkMoran's I index PSTI in the form of local spatio-temporal streamsk，
Figure BDA0001857041510000031
Wherein n is the number of unit cells, wt-k,tThe weight value is the time weight value from the t-k moment to the t moment; w is ai,jIs an element of a spatial weight matrix W, which is a row normalized spatial weight matrix used to quantify the proximity between surrounding regions; a. thei,t-kThe normalized attribute value of the cellular unit i at the time t-k is obtained;
Figure BDA0001857041510000032
the average value of the standardized attribute values of all the cellular units at the t-k moment is taken as the average value; a. thej,tThe normalized attribute value of the cellular unit j at the time t;
Figure BDA0001857041510000033
the average value of the standardized attribute values of all the cellular units at the time t is obtained; moran's I index OSTI based on geographical multivariate flow datakAnd PSTIkThe range of variation of (1) is (-1, 1).
Further, step 10 includes obtaining the OSTIkAnd PSTIkOn the basis of the value, constructing a space-time statistic Z obeying normal distribution, and testing the significance of space-time autocorrelation according to the space-time statistic Z, wherein the space-time statistic Z is expressed as:
Figure BDA0001857041510000034
or
Figure BDA0001857041510000035
Wherein:
Figure BDA0001857041510000036
Figure BDA0001857041510000041
the invention has the beneficial effects that: the analysis of the cellular unit extracts various influence factors, can more accurately obtain the cellular automaton model parameters, has smaller error than the manually determined parameters in use, and is accurate and high in efficiency; the conversion rule obtained by the ANN algorithm has dynamic property compared with the fixed conversion rule of the whole model, and can describe and accord with the actual conversion condition of the cells; the correlation among the cells is expressed according to Moran's I, the space-time distribution condition of the geographic data is better and more clearly reflected, and therefore the subsequent space-time data model simulation and prediction are more conveniently carried out, and the simulation and prediction precision is higher.
Drawings
FIG. 1 is a schematic diagram of a standard cellular automata model.
Fig. 2 is a flow chart of an embodiment.
FIG. 3 is a diagram of three cell neighborhood relationships.
FIG. 4 is a schematic diagram of model parameter acquisition by neural network training.
Detailed Description
The technical solution of the present invention can be implemented by a person skilled in the art using computer software technology. The embodiment of the invention provides the concrete steps of the embodiment of the invention by taking the land utilization data as an example and combining the accompanying figure 2, and the embodiment is described as follows:
(1) analyzing the diversity and the time-space characteristics of data by adopting geographical multivariate stream data, wherein the data adopted in the example is multi-period land utilization change remote sensing image data;
(2) preprocessing the original data, and classifying images by using ArcGIS software to obtain grid data containing various land types, such as urban land, forest land, grassland, garden land, wetland, water body land types, undeveloped land and the like;
(3) and constructing a cellular automaton model, improving the cellular automaton model by combining example data, and establishing the cellular automaton model which accords with the geographic multi-source flow data characteristics. The standard cellular automaton is a quadruple (as shown in fig. 1) composed of (cells, cell states, neighborhoods, and state update rules), and can be expressed as:
a ═ { L, d, S, N, f } equation (1)
The complete expression and description of the geographical multivariate flow data need to meet the requirement of having rich and accurate semantics, and the traditional cellular automata ignores the conversion rule and the spatial heterogeneity of asynchronous evolution of different cells (geographical regions). The cellular automaton model based on the geographic multivariate flow data is constructed by improving the cellular automaton model and is used for expressing complex, dynamic and interconnected geographic information primitives. The improved model is defined as follows:
a ═ { F, L, G, S, N, R } equation (2)
In the above formula, F ═ { F } represents attribute information data that can be expressed in a quantized manner; l ═ { L } represents the distribution of geographic phenomena, which cannot be expressed quantitatively; g is the cellular partitioning of the geospatial space; n ═ c1,c2,...,cnIs a cellular neighbor; s ═ ScRepresents a state variable of the cell; r ═ { R } represents a set of rules, otherwise known as a set of evolution functions.
According to the time series model principle, the state of the cell m at the t + k moment is determined by the state of the cell at the t moment and the cell neighbors, and the iterative process is expressed as follows:
st+k＝r(Nm,st) Formula (3)
(4) Defining a cell unit, and defining each grid in the grid data as a cell unit;
(5) multiple attribute characteristics (namely influence factors) exist in each cellular unit, and the multiple influence factors (x) are extracted by using a space analysis tool in ArcGIS softwaret,1,xt,2......,xt,Q) Q is the number of influence factors, and aiming at land utilization data, the influence factors are mainly extracted, such as distance data from a city center, distance data from a road, the number of statistical units extracted by various land types in a neighborhood window form, and the like;
(6) and determining the state condition of the cells, and representing the land type by using the state of the cells. In the image classification, the multi-color rendering is adopted to represent different land types, so that the land and land type distribution condition can be clearly known;
(7) determining neighborhood relations among unit cells, wherein currently mainstream neighborhood relation models mainly comprise von Neumann type, Moore type, Magoless type and the like, and according to example data, a Moore neighborhood relation model is adopted (as shown in figure 3);
(8) and randomly extracting attribute data in a plurality of cells as training sample data. The training data is mainly to randomly extract each grid data and various influence factor data therein. Considering that the obtained training data are different in size and have large numerical difference, the maximum value and the minimum value are adopted for standardization, so that the obtained training sample data are all between 0 and 1;
(9) repeatedly training the training sample data by using an ANN algorithm in machine learning by considering the diversity and the space-time property of the data until an optimal solution is obtained, and taking the obtained output result as a parameter required by a cellular automaton model (as shown in FIG. 4);
(10) the extracted influence factor (x)t,1,xt,2......,xt,Q) As an input layer, adopting a result obtained by training as an initial weight value;
(11) and excavating a conversion rule in each cell by using the same ANN algorithm again, entering a hidden layer through information and weight values of an input layer, obtaining an initial conversion probability of converting the cell land type into another land type through activation of a Sigmoid function, selecting a certain threshold value, returning to the input layer, and performing repeated training until the optimal conversion probability is obtained. The activation Sigmoid function selected in the ANN algorithm is as follows:
Figure BDA0001857041510000051
in the formula, wp,qRepresents the weight value between neurons p and q; x is the number oft,qRepresenting the normalized attribute value of the qth neuron (i.e., the influencer) at time t.
(12) Extracting the maximum value by using output layer data obtained by training, namely conversion probability data, determining whether the cellular land type changes at the next moment, and determining the trend of which land type is converted, wherein the calculation mode is as follows:
pt(i,j)＝(1+(-lnγ)a)×pg×con(St(i,j))×Ωt(i, j) formula (5)
In the formula, pt(i, j) is the conversion probability of the land class represented by a certain cell; 1+ (-ln γ)aRandom items are expressed, so that the simulation result is more consistent with the actual situation; p is a radical ofgIs the global transition probability; con (S)t(i, j))) represents a constraint condition for a cell unit; omegat(i, j) represents a neighborhood function representing the effect of a neighborhood on the cell transition probability。
(13) Calculating the space-time weight value w between cells by adopting a Moore neighborhood mode of a cell modelt-k,t，wi,j。wt-k,tThe weight value is the time weight value from the t-k moment to the t moment; w is ai,jIs an element of the spatial weight matrix W. W is a row-normalized spatial weight matrix used to quantify the proximity between surrounding regions.
(14) According to the continuous characteristics of the geographic multivariate stream data and the model expression form on three levels of spatial position, attribute and time, the Spatio-Temporal Moran's I (Spatio-Temporal Moran's I, STI) proposed by Wartenberg scholars is improvedk) And analyzing the space-time structure of the variable, and performing space-time autocorrelation analysis on the geographical multivariate flow data. When the analysis is performed by considering the global and local regions, the spatial and temporal heterogeneity exists, and there are two expression modes: moran's I index OSTI in the form of a global spatio-temporal streamkMoran's I index PSTI in the form of local spatio-temporal streamsk. Moran's I index OSTI in the form of spatio-temporal streamskAnd PSTIkThe expression is as follows:
Figure BDA0001857041510000061
in the formula, n is the number of the unit cell; a. thei,t-kThe normalized attribute value of the cellular unit i at the time t-k is obtained;
Figure BDA0001857041510000062
the average value of the standardized attribute values of all the cellular units at the t-k moment is taken as the average value; a. thej,tThe normalized attribute value of the cellular unit j at the time t;
Figure BDA0001857041510000063
is the average of the normalized property values of all the cellular units at time t.
Moran's I index OSTI based on geographical multivariate flow datakAnd PSTIkThe range of variation of (1) is (-1, 1). STI if two cell units are uncorrelated in space-timek(OSTIkAnd PSTIk) Is expected to be close to 0 when STIk(OSTIkAnd PSTIk) When the value of (a) is a negative value, the time-space negative autocorrelation is generally expressed, namely the time-space autocorrelation of the terrain change at the next time is weak; when STIk(OSTIkAnd PSTIk) When the value of (b) is a positive value, it generally indicates that the spatio-temporal autocorrelation, i.e., the spatio-temporal correlation of the terrain change at the next time is strong.
In obtaining the theoretical STIkOn a value basis, a spatio-temporal statistic Z obeying normal distribution can be constructed to test the significance (p) of spatio-temporal autocorrelation<0.05). The spatio-temporal statistic Z is represented as:
Figure BDA0001857041510000071
wherein:
Figure BDA0001857041510000072
for the same reason, OSTIkAnd PSTIkThe significance of the spatio-temporal autocorrelation can also be examined by equations 7 and 8.
(15) Finally, considering the correlation between the cells, the spatiotemporal weighting matrix W calculated in the step (13) and the spatiotemporal Moran's I exponential OSTI in the step (14) are combinedkOr PSTIkAnd performing simulation analysis and prediction by combining a cellular automaton model established based on land utilization data to form a combined model with a space-time autocorrelation analysis function on the basis of the cellular automaton model. The analysis and prediction result shows that non-urban land (grassland, forest land, water body and the like) is converted into urban land along with time, the change area can be counted, and the method has good guiding significance for subsequent related departments to predict the land utilization change condition in urban expansion or expansion construction.
The specific embodiments described herein are merely illustrative of the spirit of the invention. Various modifications or additions may be made to the described embodiments or alternatives may be employed by those skilled in the art without departing from the spirit or ambit of the invention as defined in the appended claims.

Claims (4)

1. A geographical multivariate stream data space-time autocorrelation analysis method based on a cellular automaton is characterized by comprising the following steps:
step 1, classifying original geographical multivariate stream data to obtain raster data containing various categories;
step 2, improving a cellular automaton model, and establishing the cellular automaton model which accords with the characteristics of geographical multivariate stream data;
the cellular automata model conforming to the characteristics of the geographical multivariate flow data in step 2 is defined as follows,
a ═ { F, L, G, S, N, R } equation (2)
In the above formula, F ═ { F } represents attribute information data that can be expressed in a quantized manner; l ═ { L } represents the distribution of geographic phenomena; g is the cellular partitioning of the geospatial space; n ═ c1,c2,...,cnIs a cellular neighbor; s ═ ScRepresents a state variable of the cell; r ═ { R } represents a set of rules, otherwise known as a set of evolution functions;
step 3, defining each grid in the grid data as a cell unit, and extracting an influence factor in each cell unit;
step 4, determining the state condition of the cells and the neighborhood relationship among the cells;
step 5, randomly extracting attribute data in a plurality of cells as training sample data;
step 6, repeatedly training the training sample data by using an ANN algorithm in machine learning by considering the diversity and the space-time property of the data until an optimal solution is obtained, and taking the obtained output result as a parameter required by the cellular automaton model;
step 7, taking the extracted influence factors as an input layer, adopting the result obtained by training as an initial weight value, and mining a conversion rule in each cell by using an ANN algorithm to obtain an initial conversion probability of converting the cell type into another type;
step 8, extracting the maximum value by using the output layer data obtained by training, namely the conversion probability data, determining whether the cell type changes at the next moment, and determining the trend of the type of conversion;
step 9, calculating a space-time weight matrix between the cells by adopting a molar neighborhood mode of the cell model;
step 10, improving a space-time Moran's I analysis variable space-time structure according to the continuous characteristics of the geographical multivariate stream data and the model expression forms on three levels of space position, attribute and time to obtain an improved Moran's I index;
the Moran's I index improved in the step 10 has two expression modes, the specific expression is as follows,
moran's I index OSTI in the form of a global spatio-temporal streamkMoran's I index PSTI in the form of local spatio-temporal streamsk，
Figure FDA0002892257290000021
Wherein n is the number of unit cells, wt-k,tThe weight value is the time weight value from the t-k moment to the t moment; w is ai,jIs an element of a spatial weight matrix W, which is a row normalized spatial weight matrix used to quantify the proximity between surrounding regions; a. thei,t-kThe normalized attribute value of the cellular unit i at the time t-k is obtained;
Figure FDA0002892257290000022
the average value of the standardized attribute values of all the cellular units at the t-k moment is taken as the average value; a. thej,tThe normalized attribute value of the cellular unit j at the time t;
Figure FDA0002892257290000023
the average value of the standardized attribute values of all the cellular units at the time t is obtained; a. thei,tThe normalized attribute value of the cellular unit i at the time t is shown;
Figure FDA0002892257290000024
the average value of the standardized attribute values of all the cellular units at the time t is obtained; moran's I index OSTI based on geographical multivariate flow datakAnd PSTIkThe variation range of (1) is (-1, 1);
and 11, considering the correlation among the cells, combining the space-time weight matrix calculated in the step 9 and the Moran's I index improved in the step 10 with a cellular automaton model established based on geographical multivariate stream data to perform simulation analysis and prediction to form a combined model with a space-time autocorrelation analysis function on the basis of the cellular automaton model.
2. The method for spatiotemporal autocorrelation analysis of geographical multivariate data based on cellular automata as claimed in claim 1, characterized in that: and in the step 5, the maximum value and the minimum value are adopted for standardization, so that the training sample data obtained by the space-time weight is between 0 and 1.
3. The method for spatiotemporal autocorrelation analysis of geographical multivariate data based on cellular automata as claimed in claim 1, characterized in that: in step 8, it is determined whether the cell type is changed at the next time, and the calculation formula of the trend of determining which type is converted is as follows,
pt(x,y)＝(1+(-lnγ)a)×pg×con(St(x,y))×Ωt(x, y) formula (5)
In the formula, pt(x, y) is the transition probability of the class represented by a certain cell; 1+ (-ln γ)aRandom items are expressed, so that the simulation result is more consistent with the actual situation; p is a radical ofgIs the global transition probability; con (S)t(x, y))) represents a constraint condition for a unit cell; omegat(x, y) represents a neighborhood function, representing the effect of a neighborhood on the cell transition probability.
4. The method for spatiotemporal autocorrelation analysis of geographical multivariate data based on cellular automata as claimed in claim 1, characterized in that: step 10 also includes obtaining the OSTIkAnd PSTIkOn the basis of the value, constructing a space-time statistic Z obeying normal distribution, and testing the significance of space-time autocorrelation according to the space-time statistic Z, wherein the space-time statistic Z is expressed as:
Figure FDA0002892257290000031
or
Figure FDA0002892257290000032
Wherein:
Figure FDA0002892257290000033
Figure FDA0002892257290000034
CN201811318953.7A 2018-11-07 2018-11-07 Geographical multivariate stream data space-time autocorrelation analysis method based on cellular automaton Active CN109508360B (en)

Priority Applications (2)

Application Number Priority Date Filing Date Title
CN201811318953.7A CN109508360B (en) 2018-11-07 2018-11-07 Geographical multivariate stream data space-time autocorrelation analysis method based on cellular automaton
AU2018101946A AU2018101946A4 (en) 2018-11-07 2018-12-04 Geographical multivariate flow data spatio-temporal autocorrelation analysis method based on cellular automaton

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811318953.7A CN109508360B (en) 2018-11-07 2018-11-07 Geographical multivariate stream data space-time autocorrelation analysis method based on cellular automaton

Publications (2)

Publication Number Publication Date
CN109508360A CN109508360A (en) 2019-03-22
CN109508360B true CN109508360B (en) 2021-03-16

Family

ID=65009405

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811318953.7A Active CN109508360B (en) 2018-11-07 2018-11-07 Geographical multivariate stream data space-time autocorrelation analysis method based on cellular automaton

Country Status (2)

Country Link
CN (1) CN109508360B (en)
AU (1) AU2018101946A4 (en)

Families Citing this family (15)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110909924B (en) * 2019-11-12 2023-05-02 同济大学 Urban expansion multi-scenario simulation cellular automaton method based on cross entropy optimizer
CN111080070B (en) * 2019-11-19 2023-05-02 同济大学 Urban land utilization cellular automaton simulation method based on space errors
CN111428405B (en) * 2020-03-20 2023-07-07 北京百分点科技集团股份有限公司 Fine particulate matter concentration simulation method and device, storage medium and electronic equipment
CN111797523B (en) * 2020-06-30 2024-03-22 南京图申图信息科技有限公司 Land utilization change-oriented multi-level vector cellular automaton modeling method
CN112396231B (en) * 2020-11-18 2023-09-29 京东科技控股股份有限公司 Modeling method and device for space-time data, electronic equipment and readable medium
CN113032510B (en) * 2021-04-02 2023-06-02 中国人民解放军国防科技大学 LBGM method initial value disturbance generation method based on Gaussian weight
CN113297174B (en) * 2021-05-24 2023-10-13 中南大学 Land utilization change simulation method based on deep learning
CN113255221B (en) * 2021-05-28 2022-11-25 中国石油大学(华东) Inter-well connectivity analysis method and device based on cellular automata deduction
CN113742925B (en) * 2021-09-06 2023-09-29 吉林大学 Signal control intersection pedestrian flow simulation method
CN113793376B (en) * 2021-09-14 2023-12-05 安徽理工大学 Irrigation water body extraction method
CN114611188B (en) * 2022-03-09 2022-09-02 生态环境部卫星环境应用中心 Cellular automata-based tailing pond dam break leakage simulation analysis method and system
CN116451873B (en) * 2023-06-12 2023-10-20 中国科学技术大学 Wind power generation power prediction method and system based on multi-scale double space-time network area
CN116955335A (en) * 2023-07-21 2023-10-27 北京国信达数据技术有限公司 Address data management method and system based on big data model algorithm
CN116976526B (en) * 2023-09-20 2023-12-15 北京师范大学 Land utilization change prediction method coupling ViViViT and ANN
CN117057674A (en) * 2023-10-11 2023-11-14 晨达(广州)网络科技有限公司 Discrete model-based information technology service quality assessment method and system

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102184328A (en) * 2011-05-10 2011-09-14 南京大学 Method for optimizing land use evolution CA model transformation rules
CN104732279A (en) * 2015-03-25 2015-06-24 武汉大学 Improved cellular automaton traffic flow simulation analysis method based on geographic information system
CN106780089A (en) * 2016-12-27 2017-05-31 广州蓝图地理信息技术有限公司 Permanent basic farmland demarcation method based on neutral net cellular Automation Model
CN107194504A (en) * 2017-05-09 2017-09-22 云南师范大学 Forecasting Methodology, the device and system of land use state

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8392166B2 (en) * 2007-09-22 2013-03-05 Chaitanya N. Hiremath Method for providing a numerical and graphical abbreviated profile of a drug (A-POD)

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102184328A (en) * 2011-05-10 2011-09-14 南京大学 Method for optimizing land use evolution CA model transformation rules
CN104732279A (en) * 2015-03-25 2015-06-24 武汉大学 Improved cellular automaton traffic flow simulation analysis method based on geographic information system
CN106780089A (en) * 2016-12-27 2017-05-31 广州蓝图地理信息技术有限公司 Permanent basic farmland demarcation method based on neutral net cellular Automation Model
CN107194504A (en) * 2017-05-09 2017-09-22 云南师范大学 Forecasting Methodology, the device and system of land use state

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
基于GIS和元胞自动机的女儿寨小流域植被动态模拟研究;赵占轻;《中国优秀硕士学位论文全文数据库 基础科学辑》;20111215;A006-272 *

Also Published As

Publication number Publication date
CN109508360A (en) 2019-03-22
AU2018101946A4 (en) 2019-01-17

Similar Documents

Publication Publication Date Title
CN109508360B (en) Geographical multivariate stream data space-time autocorrelation analysis method based on cellular automaton
CN111161535B (en) Attention mechanism-based graph neural network traffic flow prediction method and system
CN109492822B (en) Air pollutant concentration time-space domain correlation prediction method
CN111860951A (en) Rail transit passenger flow prediction method based on dynamic hypergraph convolutional network
CN111339818B (en) Face multi-attribute recognition system
CN110138595A (en) Time link prediction technique, device, equipment and the medium of dynamic weighting network
CN114092697B (en) Building facade semantic segmentation method with attention fused with global and local depth features
CN111079999A (en) Flood disaster susceptibility prediction method based on CNN and SVM
CN112766280A (en) Remote sensing image road extraction method based on graph convolution
WO2022257190A1 (en) Quantum walk-based multi-feature simulation method for behavior trajectory sequences
CN114693064B (en) Building group scheme generation performance evaluation method
CN114944053A (en) Traffic flow prediction method based on spatio-temporal hypergraph neural network
Xu et al. A graph spatial-temporal model for predicting population density of key areas
CN107665376A (en) A kind of Wetland Space changes in distribution framework analogue and Forecasting Methodology
Dai et al. Spatio-temporal deep learning framework for traffic speed forecasting in IoT
CN114519302A (en) Road traffic situation simulation method based on digital twin
Ayhan et al. Analysis of image classification methods for remote sensing
Chou et al. Imaging time-series with features to enable visual recognition of regional energy consumption by bio-inspired optimization of deep learning
CN115629160A (en) Air pollutant concentration prediction method and system based on space-time diagram
CN116187835A (en) Data-driven-based method and system for estimating theoretical line loss interval of transformer area
Zhang et al. Urban spatial risk prediction and optimization analysis of POI based on deep learning from the perspective of an epidemic
CN113326862A (en) Audit big data fusion clustering and risk data detection method, medium and equipment
CN116662468A (en) Urban functional area identification method and system based on geographic object space mode characteristics
Zhai et al. Self-supervision spatiotemporal part-whole convolutional neural network for traffic prediction
CN111209968A (en) Multi-meteorological factor mode forecast temperature correction method and system based on deep learning

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant