CN111242394A - Method and system for extracting spatial correlation characteristics - Google Patents
Method and system for extracting spatial correlation characteristics Download PDFInfo
- Publication number
- CN111242394A CN111242394A CN202010336499.9A CN202010336499A CN111242394A CN 111242394 A CN111242394 A CN 111242394A CN 202010336499 A CN202010336499 A CN 202010336499A CN 111242394 A CN111242394 A CN 111242394A
- Authority
- CN
- China
- Prior art keywords
- data set
- spatial correlation
- extracting
- dependence
- periodic
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 23
- 238000000605 extraction Methods 0.000 claims abstract description 54
- 230000000737 periodic effect Effects 0.000 claims abstract description 43
- 238000012545 processing Methods 0.000 claims abstract description 37
- 238000010586 diagram Methods 0.000 claims abstract description 26
- 230000007246 mechanism Effects 0.000 claims abstract description 23
- 238000004220 aggregation Methods 0.000 claims abstract description 19
- 230000002776 aggregation Effects 0.000 claims abstract description 19
- 230000003442 weekly effect Effects 0.000 claims description 22
- 238000010276 construction Methods 0.000 claims description 12
- 230000004931 aggregating effect Effects 0.000 claims description 7
- 230000001419 dependent effect Effects 0.000 claims description 7
- 230000002596 correlated effect Effects 0.000 claims description 5
- 239000011159 matrix material Substances 0.000 description 38
- 238000013527 convolutional neural network Methods 0.000 description 7
- 230000006870 function Effects 0.000 description 7
- 230000004913 activation Effects 0.000 description 6
- 238000013528 artificial neural network Methods 0.000 description 6
- 230000000875 corresponding effect Effects 0.000 description 5
- 238000012549 training Methods 0.000 description 5
- 230000007613 environmental effect Effects 0.000 description 4
- 239000000284 extract Substances 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 230000000694 effects Effects 0.000 description 3
- 230000000306 recurrent effect Effects 0.000 description 3
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 230000002452 interceptive effect Effects 0.000 description 2
- 230000007787 long-term memory Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000000926 separation method Methods 0.000 description 2
- 230000036962 time dependent Effects 0.000 description 2
- NAWXUBYGYWOOIX-SFHVURJKSA-N (2s)-2-[[4-[2-(2,4-diaminoquinazolin-6-yl)ethyl]benzoyl]amino]-4-methylidenepentanedioic acid Chemical compound C1=CC2=NC(N)=NC(N)=C2C=C1CCC1=CC=C(C(=O)N[C@@H](CC(=C)C(O)=O)C(O)=O)C=C1 NAWXUBYGYWOOIX-SFHVURJKSA-N 0.000 description 1
- 238000012935 Averaging Methods 0.000 description 1
- 102100033592 Calponin-3 Human genes 0.000 description 1
- 101000945410 Homo sapiens Calponin-3 Proteins 0.000 description 1
- 101001095088 Homo sapiens Melanoma antigen preferentially expressed in tumors Proteins 0.000 description 1
- 102100037020 Melanoma antigen preferentially expressed in tumors Human genes 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 230000005012 migration Effects 0.000 description 1
- 238000013508 migration Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001360 synchronised effect Effects 0.000 description 1
Images
Classifications
-
- 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
- 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
- G06N3/044—Recurrent networks, e.g. Hopfield networks
-
- 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
- G06N3/045—Combinations of networks
-
- 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
- G06N3/08—Learning methods
-
- 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
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Economics (AREA)
- Evolutionary Computation (AREA)
- Computational Linguistics (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Biophysics (AREA)
- Mathematical Physics (AREA)
- Biomedical Technology (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Tourism & Hospitality (AREA)
- Data Mining & Analysis (AREA)
- Strategic Management (AREA)
- Human Resources & Organizations (AREA)
- Marketing (AREA)
- Development Economics (AREA)
- General Business, Economics & Management (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Educational Administration (AREA)
- Primary Health Care (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention belongs to the technical field of data processing, and discloses a method and a system for extracting spatial correlation characteristics, wherein the method for extracting the spatial correlation characteristics comprises the following steps: step S1: selecting the spatial correlation data to obtain at least one periodic data set and a trend data set; step S2: processing at least one periodic data set based on a time sliding mechanism, and performing feature extraction and aggregation on the processed at least one periodic data set to obtain a feature map; step S3: and constructing a residual convolution unit, and extracting spatial correlation characteristics according to the trend data set and the characteristic diagram. The invention realizes accurate description of complex space dependence and provides more accurate data support for space-time prediction in the future.
Description
Technical Field
The invention belongs to the field of data processing, and particularly relates to a method and a system for extracting spatial correlation characteristics.
Background
The OD data refers to the magnitude of a social or economic interactive relationship variable from a departure point (o point) to a destination point (d point) in a certain period of time, and is also called OD data. The interactive relationship between the spatial geographic units in a certain time period comprises the economic trade scale, the social communication frequency or the passenger and goods transportation volume between areas. The combination of the OD quantities between all the origin (o) and destination (d) points is called the OD matrix. Prediction of OD data is often an important basis for making socio-economic activity plans and decisions between regions. Generally, the task of OD data prediction is to predict OD data in a subsequent period on the premise of given historical OD data (and corresponding environmental information), so as to be an important basis for socio-economic activity organization and management decision.
For a certain traffic line network, there exists a complex spatial dependency relationship of OD data in space, and such spatial dependency can be further divided into origin dependency and destination dependency according to cause. The two spatial dependencies differ significantly from the spatial dependency relationship embodied in conventional spatial data. How to accurately identify the data is an important basis for accurate prediction of OD data.
The origin depends on: there is a correlation between OD amounts starting from an adjacent departure point and ending at a specific destination. For example, in a subway line network, the same change law is exhibited by the traveling amounts of passengers arriving at the same station in the center of a city, starting from adjacent stations in the vicinity of a certain residential area.
Destination dependence: there is a correlation between OD amounts starting from a specific departure point and ending at an adjacent destination point. For example, in a subway line network, the traveling volume of passengers departing from a station in a certain residential area and arriving at an adjacent station in the center of a city has a strong positive correlation.
In recent years, successful application of deep learning in various fields has stimulated research and attempts for its application in the fields of transportation and transportation. For example, some studies consider road network traffic throughout a city as a thermodynamic diagram (where each pixel value represents traffic within the corresponding region), and model the spatial dependence of the non-linearity using Convolutional Neural Networks (CNNs). In addition, some researchers have proposed using Recurrent Neural Networks (RNNs) to build non-linear time-dependent models for traffic flow prediction. In subsequent researches, CNN and RNN are organically fused, a comprehensive prediction model considering the time and space dependence relationship is provided, and the prediction accuracy is further improved. However, in many of these studies, urban road traffic flow is considered as a study target, and the spatial distribution and the dependency characteristics that require the OD data in a complicated manner are not considered enough, and it is difficult to realize accurate prediction.
Disclosure of Invention
In view of the above problem, the present invention provides a method for extracting spatial correlation features, where the method includes:
step S1: selecting the spatial correlation data to obtain a periodic data set and a trend data set;
step S2: processing the periodic data set based on a time sliding mechanism, and then performing feature extraction and aggregation on the periodic data set to obtain a feature map;
step S3: and constructing a residual convolution unit, and extracting spatial correlation characteristics according to the trend data set and the characteristic diagram.
In the above extraction method, step S1 includes:
and selecting a periodic data set and a trend data set by taking different time periods as prediction targets according to the spatial correlation data.
In the above extraction method, the cycle data set includes a weekly interval historical cycle data set and a daily interval historical cycle data set, and the trend data set is a current day trend data set.
In the above extraction method, step S2 includes:
step S21: processing the weekly interval historical period data set and the daily interval historical period data set based on a time sliding mechanism to obtain a sliding weekly interval historical period data set and sliding daily interval historical period data;
step S22: and performing feature extraction and aggregation on the sliding week interval historical period data set and the sliding day interval historical period data set through a two-dimensional point-by-point convolution layer to obtain a feature map.
In the above extraction method, step S3 includes:
step S31: extracting the characteristics of departure place dependence and destination dependence by stacking a plurality of depth-separated one-dimensional convolution layers;
step S32: and extracting a spatial correlation structure through the one-dimensional point-by-point convolutional layer, and aggregating the departure place dependence and destination dependence characteristics to output spatial correlation characteristics.
The invention also provides a system for extracting the spatial correlation characteristics, which comprises the following steps:
a data set construction unit: selecting the spatial correlation data to obtain a periodic data set and a trend data set;
a feature map obtaining unit: processing the periodic data set based on a time sliding mechanism, and then performing feature extraction and aggregation on the periodic data set to obtain a feature map;
residual convolution unit construction unit: and constructing a residual convolution unit, and extracting spatial correlation characteristics according to the trend data set and the characteristic diagram.
In the above extraction system, the data set construction unit selects the cycle data set and the trend data set for the prediction target at different time periods according to the spatial correlation data.
In the above extraction system, the cycle data set includes a weekly interval historical cycle data set and a daily interval historical cycle data set, and the trend data set is a current day trend data set.
In the above extraction system, the feature map obtaining unit includes:
a data set processing module: processing the weekly interval historical period data set and the daily interval historical period data set based on a time sliding mechanism to obtain a sliding weekly interval historical period data set and sliding daily interval historical period data;
a characteristic diagram obtaining module: and performing feature extraction and aggregation on the sliding week interval historical period data set and the sliding day interval historical period data set through a two-dimensional point-by-point convolution layer to obtain a feature map.
In the above extraction system, the residual convolution unit construction unit includes:
a dependent feature extraction module: extracting the characteristics of departure place dependence and destination dependence by stacking a plurality of depth-separated one-dimensional convolution layers;
the spatial correlation characteristic output module: and extracting a spatial correlation structure through the one-dimensional point-by-point convolutional layer, and aggregating the departure place dependence and destination dependence characteristics to output spatial correlation characteristics.
Aiming at the prior art, the invention has the following effects: aiming at the complex spatial correlation of spatial correlation data, a passenger flow residual error convolution unit is constructed, meanwhile, departure place dependence and destination dependence characteristics are identified, accurate description of the complex spatial dependence is achieved, and more accurate data support is provided for space-time prediction in the future.
Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.
FIG. 1 is a flow chart of the extraction method of the present invention;
FIG. 2 is a flowchart illustrating the substeps of step S2 in FIG. 1;
FIG. 3 is a flowchart illustrating the substeps of step S3 in FIG. 1;
FIG. 4 is a schematic diagram of the extraction system;
FIG. 5 is a schematic diagram of an OD sparse spatiotemporal residual network model;
FIG. 6 is a schematic diagram of a sliding time window mechanism for periodic historical feature extraction;
FIG. 7 is a schematic diagram of the structure of an OD residual convolution unit;
FIG. 8 is a schematic diagram of spatial feature extraction;
FIG. 9 is a simplified schematic diagram of a sequential processing unit;
FIG. 10 is a schematic diagram of a sequential stacked feature matrix processing procedure;
fig. 11 is a schematic diagram of a non-zero element attention mechanism considering sparsity.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As used herein, the terms "comprising," "including," "having," "containing," and the like are open-ended terms that mean including, but not limited to.
References to "a plurality" herein include "two" and "more than two".
Referring to fig. 1, fig. 1 is a flow chart of the extraction method of the present invention. As shown in fig. 1, the method for extracting spatial correlation features of the present invention includes:
step S1: and selecting the spatial correlation data to obtain at least one periodic data set and a trend data set.
Specifically, in step S1, at least one period data set and a trend data set are selected from the spatial correlation data and the prediction targets in different time periods, wherein the at least one period data set includes at least one of a week interval historical period data set, a day interval historical period data set, a month interval historical period data set and a year interval historical period data set, and the trend data set is a current day trend data set.
Step S2: and processing at least one periodic data set based on a time sliding mechanism, and performing feature extraction and aggregation on the processed at least one periodic data set to obtain a feature map.
Referring to fig. 2, fig. 2 is a flowchart illustrating a sub-step of step S2 in fig. 1. As shown in fig. 2, the step S2 includes:
step S21: processing at least one periodic data set based on a time sliding mechanism to correspondingly obtain at least one sliding periodic data set;
step S22: and performing feature extraction and aggregation on at least one sliding period data set through a two-dimensional point-by-point convolution layer to obtain a feature map.
Step S3: and constructing a residual convolution unit, and extracting spatial correlation characteristics according to the trend data set and the characteristic diagram.
Referring to fig. 3, fig. 3 is a flowchart illustrating a substep of step S3 in fig. 1. As shown in fig. 3, the step S3 includes:
step S31: extracting the characteristics of departure place dependence and destination dependence by stacking a plurality of depth-separated one-dimensional convolution layers;
step S32: and extracting a spatial correlation structure through the one-dimensional point-by-point convolutional layer, and aggregating the departure place dependence and destination dependence characteristics to output spatial correlation characteristics.
Referring to fig. 5-10, the principle of the extraction method will be described with reference to fig. 5-10 in an embodiment. It should be noted that, in the embodiment, the extraction method of the present invention is applied to demand prediction of rail transit passenger flow, but the present invention is not limited thereto, and in other embodiments, the present invention can also be applied to analysis and prediction of other socioeconomic phenomena having origin-destination points at the same time, such as trade between countries and regions, population migration, and various types of passenger and cargo transportation.
First, the spatial correlation data is preprocessed.
Specifically, the weekly interval historical cycle data set, the daily interval historical cycle data set, and the current day trend data set are selected from the historical data using different time periods as prediction targets in this step.
Definition 1: for a railway transit network including N stations, the set of all stations is S (S = {1,2,3, …, N }), and the set of all trips in the t-th time slot of the study is PtFor stations i (i ∈ S), j (j ∈ S), OD passenger flow volume from station i to station j in t-th time periodCan be defined as:
wherein p isoAnd pdRepresents the origin and destination of the p-th trip in the t-th time period, | circuitry represents the aggregate potential, which is the number of elements in the aggregate, in equation (1),
|{po=i∧pd= j } | denotes the set { po=i∧pdNumber of elements in = j }.
Define 2 (OD traffic matrix): in the studied railway traffic network, the OD passenger flow between all stations in the t-th time periodForming an OD passenger flow matrix X of the rail transit line network in the t-th time periodt,Xt∈RN×NR is a real number, i.e. XtThe matrix is a real number matrix of dimension N × N, as shown in the following equation.
For a particular rail transit network, the OD passenger flow matrix { X over the first m-1 time periods is giventI t =1,2,3, …, m-1}, and predicting OD traffic matrix X in the m-th time period thereofmThe problem of predicting passenger flow of a rail transit network can be expressed as follows:
in which Ω(s) are prediction models or prediction functions, EtThe environmental variables in the t-th time period are used for describing environmental information such as weather conditions (sunny days, cloudy days, rain, snow and fog), air temperature, wind speed, whether the weather is a holiday or not and the like. And secondly, constructing an OD passenger flow sparse space-time residual error network model according to the processed data.
Wherein, the method comprises the following steps:
1) and processing the weekly interval historical period data set and the daily interval historical period data set based on a time sliding mechanism to obtain a sliding weekly interval historical period data set and sliding daily interval historical period data.
Specifically, referring to fig. 5, fig. 5 is a schematic diagram of a sparse space-time residual error network model for OD passenger flow. As shown in FIG. 5, the predicted OD traffic matrix of the t-th time period is not all strongly correlated with the OD traffic matrix of the first t-1 time periods. For example, the OD passenger flow of 8:00 of the early peak of a certain Tuesday is only strongly correlated with 7:00-8:00 of the day, and is also strongly correlated with historical data around a plurality of working days and a plurality of 8:00 early peaks of the Tuesday, and is less dependent on the passenger flow data of other time periods. Therefore, the invention respectively selects a weekly interval historical period data set, a daily interval historical period data set and a current day trend data set from the historical data as input variables of the OD passenger flow sparse space-time residual error network model.
Wherein the trend data set S of the daytrFrom the top q of the pseudo-prediction period ttrThe OD traffic matrix for each time interval is constructed as shown below. Wherein,backtracking k for a quasi-prediction time period ttrThe OD traffic matrix for each time period,r is a real number, i.e. StrBelong to qtrX N dimensional real matrix.
Because the evolution law of the OD passenger flow in each period has certain difference, the OD passenger flow in the period t to be predicted has strong relevance with the observed value before the interval of the historical integer periods, and depends on the observed value of the surrounding period to a certain extent. For example, the OD traffic of 8:00 early peak on a certain tuesday is strongly correlated with the observed values of 8:00 early peak on the last tuesday and the adjacent time period. Therefore, a certain time sliding mechanism needs to be considered when constructing the weekly and daily interval historical period data sets. For the k-thwInterval of every week and kdSliding historical data set of daily intervals of individualsAndand can be represented by formula (5).
Wherein,r is a real number, i.e.Belong to (2 p)wA real matrix of + 1) x NxN dimensions,r is a real number, i.e.Belong to (2 p)dA real matrix of + 1) × nxn dimensions. q. q.swAnd q isdHistorical number of cycles in the historical cycle data set for weekly intervals and daily intervals, respectively。pwAnd pdThe time sliding window sizes are respectively. I is the number of time periods of the day, related to the length of the study period. Weekly and daily interval historical period data set SwAnd SdHistorical periodic data sets, possibly from slip cycle intervalsHistorical periodic data set spaced from sliding daysThe construction is carried out as shown in formula (6).
2) and constructing a two-dimensional point-by-point convolution layer, and performing feature extraction and aggregation on the sliding week interval historical period data set and the sliding day interval historical period data set to obtain a feature map.
Referring to fig. 6, fig. 6 is a schematic diagram of a sliding time window mechanism for periodic historical feature extraction. 5-6, part a of FIG. 5, the present invention employs two-dimensional point-by-point convolutional layer 1x1_ cov2 for sliding history data setAndcarrying out feature extraction and aggregation to obtain a feature mapAnd,,as shown in equation (7).
Wherein,andfor learnable parameters, f is an activation function,the convolution operation is shown, and the ReLU function is selected to ensure the sparsity of the feature, but the invention is not limited thereto.
And thirdly, constructing an OD passenger flow residual error convolution unit, and extracting spatial features from the feature map and the current day trend data set.
The method comprises the following steps:
1) constructing a plurality of depth separation one-dimensional convolution layers, and extracting the characteristics of departure place dependence or destination dependence by stacking the depth separation one-dimensional convolution layers;
2) and constructing a one-dimensional point-by-point convolutional layer, extracting a spatial correlation structure through the one-dimensional point-by-point convolutional layer, and aggregating the two spatial dependencies to output spatial characteristics.
Please refer to fig. 7 and 8; FIG. 7 is a schematic diagram of the OD passenger flow residual convolution unit structure; fig. 8 is a schematic diagram of spatial feature extraction. As shown in fig. 7 and 8, in part b of fig. 5, the present invention proposes an OD passenger flow residual convolution unit OD _ result for extracting the characteristics of origin dependency and destination dependency at the same time, and its structure is shown in fig. 7. OD passenger flow by stacking multiple depth-separated one-dimensional convolutional layers Ds _ cov1And the residual convolution unit extracts the departure place dependence or destination dependence characteristics, simultaneously extracts the spatial correlation structure by utilizing the one-dimensional point-by-point convolution layer 1x 1-cov 1, and outputs the two spatial dependencies in an aggregation manner. The OD passenger flow residual convolution unit also introduces residual concatenation to prevent gradient vanishing from occurring. Are respectively paired、Andextracting spatial features to obtain features、Andas shown in formula (8).
FIG. 8 shows the spatial feature extraction principle based on Ds _ cov1 and 1x1_ cov 1. Taking destination dependent feature extraction as an example, rows in the OD traffic matrix are considered as channel dimensions herein. The OD traffic matrix can be converted into a 1 xn dimensional picture containing N channels, each channel representing the OD traffic structure from a particular station. Ds _ cov1 may utilize a 1 × N structure, wherein the present invention uses a 1 × 3 convolution kernel to extract the correlation between OD traffic arriving at adjacent stations from the same station, but the present invention is not limited thereto. Because Ds _ cov1 adopts different convolution kernels in different channels, the OD passenger flow residual convolution unit realizes the destination dependent feature extraction of all departure stations. By stacking multiple Ds _ cov1 layers, the OD passenger flow residual convolution unit can further identify a larger spatial range of destination dependent features. Finally, the OD passenger flow residual convolution unit utilizes 1x1_ cov1 to identify the association structure between a certain road section (channel) and other road sections, and the description capacity of the destination association feature is further improved. The extraction process of the OD passenger flow residual convolution unit on the departure place dependence characteristic is similar to that of the destination dependence characteristic, and only the input OD passenger flow matrix needs to be transposed before extraction.
And constructing a simplified time sequence processing unit to extract nonlinear time sequence correlation characteristics from the plurality of spatial characteristics to complete time-space characteristic aggregation and obtain historical time-space prediction.
The method comprises the following steps:
1) stacking the spatial features in time to obtain a current day trend spatial feature set, a week interval period spatial feature set and a day interval period spatial feature set;
2) constructing a plurality of two-dimensional point-by-point convolution layers, and extracting nonlinear time sequence correlation characteristics from a current day trend space characteristic set, a week interval period space characteristic set and a day interval period space characteristic set through the continuous two-dimensional point-by-point convolution layers;
3) obtaining a current day trend space-time prediction value, a week interval period space-time prediction value and a day interval period space-time prediction value according to the extracted time sequence correlation characteristics;
4) and obtaining a historical space-time predicted value according to the current day trend space-time predicted value, the week interval period space-time predicted value and the day interval period space-time predicted value.
Please refer to fig. 9 and 10; FIG. 9 is a simplified schematic diagram of a sequential processing unit; FIG. 10 is a schematic diagram of a process for processing a time-series stacked feature matrix. As shown in fig. 9 and 10, a conventional Recurrent Neural Network (RNN), such as a long-term memory network (LSTM), can automatically learn long-term dependencies in a time series. However, predicting the OD passenger flow matrix of the entire traffic network requires RNN to have higher hidden layer and output layer feature dimensions, which requires a huge amount of training samples and computational resources, and is often difficult to satisfy in most application scenarios. According to the OD traffic time dependency feature described above, the OD traffic of the time period to be predicted has a strong correlation only with observations of several time periods, weeks or days before. In view of the small number of time periods in the two periodic dependencies and the daily trend dependency considered in the model, the present invention proposes a simplified timing processing unit (Simp _ SeqUnit).
Fig. 9 shows the complete structure of a simplified sequential processing unit. In the simplified time sequence processing unit, a plurality of continuous two-dimensional point-by-point convolution layers 1x1_ cov2 are used for extracting nonlinear time sequence correlation characteristics from the spatial characteristics of a plurality of time segments, and 1x1_ cov2 layers with an output channel of 1 complete final space-time characteristic aggregation. The detailed principle of the simplified sequential processing unit is shown in fig. 10. Compared with the traditional RNN, when the number of considered time periods is small, the simplified time sequence processing unit can utilize fewer parameters to realize the nonlinear time characteristic extraction and prediction of all station OD passenger flows, and is more flexible and efficient on a small data set.
Specifically, part c of fig. 5 shows the process of OD passenger flow time-dependent feature extraction using the simplified time-series processing unit. Spatial features obtained by OD passenger flow residual convolution unit、Andstacking in time to obtain a current day trend spatial feature set Utr() Periodic spatial feature set Uw() And a daily interval periodic spatial feature set Ud() As shown in formula (9).
Using reduced sequential processing unit pairs U, respectivelytr、UwAnd UdExtracting time sequence characteristics to obtain a time-space predicted value Z of the current day trendtr() And a week interval period space-time predicted value Zw() And the space-time prediction value Z of the day interval periodd() As shown in formula (10).
Wherein phi is a processing function of the simplified time sequence processing unit,、andto simplify the learnable parameters of the sequential processing unit. Summing the three space-time predicted values to obtain a historical space-time predicted value Zst,As shown in formula (11).
Then, a preliminary prediction is obtained according to the historical space-time prediction and the external environment prediction.
Specifically, the historical spatio-temporal predicted value Z is calculated under the condition of considering the influence of external environmental factorsstAnd the external environment predicted value ZEt(derived from Et by the ordinary linear layer) are summed to obtain a preliminary prediction for the t periodAs shown in formula (12).
And then, introducing a non-zero attention mechanism of sparsity to obtain final prediction according to the preliminary prediction. The method comprises the following steps:
1) obtaining an OD passenger flow matrix mean value according to the weekly interval historical period data set, the daily interval historical period data set and the current day trend data set;
2) converting the OD passenger flow matrix mean value into a non-zero element attention matrix through a non-zero activation function;
3) and filtering the preliminary predicted value through a non-zero element attention matrix to obtain a final predicted value of the OD passenger flow matrix in a certain time period.
Additionally, on the basis of current-term trend dependence, daily interval period dependence, and weekly interval period dependence considered herein, monthly interval (4 weeks) period dependence and yearly interval period dependence (52 weeks) may be further considered. In step S21 of the method proposed by the present invention, a monthly interval period and yearly interval period historical data set is added, and a processing procedure similar to the daily interval period and yearly interval period historical data set is adopted to extract corresponding spatiotemporal features, so as to be used for final OD passenger flow prediction, which may be used as another embodiment of the present invention. The calculation of the historical spatiotemporal prediction value shown in equation (11) can be updated to equation (11-1) on the basis of consideration of the monthly interval cycle dependency and the yearly interval cycle dependency, where the sums are the monthly interval cycle and yearly interval cycle spatiotemporal prediction values, respectively.
Referring to fig. 11 again, fig. 11 is a schematic diagram of a non-zero element attention mechanism considering sparsity. As shown in fig. 11, in order to describe the sparsity of the OD passenger flow in the spatial distribution, in section d of fig. 5, the present invention also introduces a Non-zero attention mechanism (Non-zero Activation), the principle of which is shown in fig. 11. The sparsity of the traffic matrix remains stable due to the OD over time. Thus, the OD passenger flow matrix mean of all input data sets is used herein() Known as sparsely distributed features. By the action of the non-zero activation function a,is converted to a non-zero element attention matrix. Use it to predict the initial valueFiltering to obtain the final predicted value of the OD passenger flow matrix in the t periodThus, the sparsity of the final prediction result is ensured, as shown in formula (13).
Where denotes the element-by-element multiplication of the matrix. As used herein, the non-zero activation function Lambda is as equation (14), and its parameter Lambda can be set according to the actual data distribution. When the parameter λ is large enough, Λ (x) approaches 1 even if the value of x is minute. It can be seen that the designed non-zero attention mechanism is not only to a certain extentTo ensureThe sparse structure of the network simultaneously limits the space-time feature extraction of the OD-SparsesSTnet network to be only carried out around non-zero elements, thereby improvingThe prediction accuracy of (2).
The OD passenger flow sparse space-time residual error network model is trained by using a back propagation rule of a deep neural network and an Adam algorithm. Specifically, a training sample set is formed according to a current day trend data set, a week interval period data set and a day interval period data set, and model training is carried out according to the training sample set by using a back propagation rule of a deep neural network and an Adam algorithm. The model training objective is to minimize the loss of the square Root of Mean Square Error (RMSE), which is shown in equation (15).
However, for the predicted sparse OD passenger flow matrix, since the general prediction deviation evaluation index, average absolute percentage error (MAPE), may have a dividend of 0, the present invention provides a total absolute percentage error (GAPE) to estimate the prediction accuracy of the entire OD passenger flow matrix, as shown in the following formula:
specifically, a set of normalized historical OD passenger flow matrix observations is input: { X1, X2, …, Xm-1 }; external environment information observation set: { E1, E2, …, Em-1 }; sequence length of weekly dependence history data, daily dependence history data and current trend history data: qw, qd, qtr weekly dependency history data, daily dependency history data sliding window: pw, pd; study period length: h (unit: minutes, total number of time periods contained in one day: I = 1440/h), an OD passenger flow sparse spatio-temporal residual network model was trained.
The invention verifies the advancement of the proposed OD passenger flow sparse space-time residual error network model by introducing 7 existing models as a reference. The basic settings of the reference model in 7 are as follows:
historical Average (HA): and directly carrying out historical synchronous averaging on the input weekly interval period data set and the daily interval period data set to serve as a predicted value of the OD passenger flow matrix at the predicted time.
Autoregressive moving average (ARIMA): autoregressive moving average is a classical model of time series prediction, which is used here to predict the time series of current day trends.
Three-dimensional convolutional network (CNN 3): the three-dimensional convolution network is directly utilized to carry out space and time feature simultaneous extraction (without independent processing of departure place and destination dependence) and prediction on the current day trend time series.
General Recurrent Neural Network (RNN): and firstly, extracting the characteristics of each OD passenger flow matrix in the time sequence of the daily trend, and then predicting by using the RNN.
Long and short term memory networks (LSTM): and (3) firstly, extracting the characteristics of each OD passenger flow matrix in the daily trend time sequence, and then predicting by using the LSTM.
Deep space-time network (deep st): the space-time deep neural network prediction model aiming at the space-time data is mainly used for carrying out people stream aggregation prediction in an urban range.
Space-time residual network (ST-ResNet): aiming at a space-time residual error network model of space-time data, a plurality of CNNs and corresponding residual error connections are stacked for space characteristic identification, and the method is mainly used for urban traffic flow prediction.
TABLE 17 comparison of predicted results of the reference model with the proposed OD-SparsesSTnet
Table 1 shows the predicted comparison of 7 reference models with the proposed OD-SparsesSTnet model over the 8:00-8:15 early peak (or 8:00-8:05 early peak). Because the HA method only partially considers the historical periodic dependence of the weekly interval and the daily interval, and the spatial dependence is not considered, the RMSE and GAPE values of the prediction result are the highest, and the prediction deviation is the largest. The ARIMA, the RNN and the LSTM are time sequence models essentially, can better identify the dependence of the trend of the day but cannot identify the dependence of the OD passenger flow in a complex space, and the prediction indexes of the ARIMA, the RNN and the LSTM are only superior to those of the HA method. Particularly, when the output characteristic dimension is large, due to the fact that a large number of hidden layer parameters are needed by the RNN and the LSTM, the method is difficult to be applied to OD passenger flow matrix prediction of the whole traffic network, and large prediction deviation exists.
The CNN3 can regard the time change of the OD passenger flow matrix as a third dimension, can extract the time-of-day dependence and space dependence characteristics to a certain extent, and the prediction result is superior to three time sequence models and HA methods. As a typical spatio-temporal data processing model, the DeepST and the ST-ResNet can simultaneously extract multiple time correlation and external factor characteristics, and the internal CNN can also identify spatial dependence to a certain extent. The prediction deviation of the two models is greatly reduced relative to other reference models. However, deep ST and ST-ResNet still cannot finely describe the dependency of origin and destination specific to OD traffic, and there is no consideration for sparse distribution of data.
The proposed OD-SparsesSTnet further introduces a non-zero activation mechanism to describe the sparse distribution characteristic of OD passenger flows on the basis of considering the complex time dependence (week interval period dependence, day interval period dependence, current day trend dependence) and space dependence (departure place dependence and destination dependence) specific to the OD passenger flows of the traffic line network. The prediction deviation is obviously lower than that of the existing space model (CNN 3), time sequence model (ARIMA, RNN and LSTM) and space-time model (DeepsT and ST-ResNe), and the prediction deviation is reduced by more than 14.89%.
Referring to fig. 4, fig. 4 is a schematic structural diagram of an extraction system. As shown in fig. 4, the extraction system of the present invention comprises:
the data set construction unit 11: selecting the spatial correlation data to obtain at least one periodic data set and a trend data set;
feature map obtaining unit 12: processing at least one periodic data set based on a time sliding mechanism, and performing feature extraction and aggregation on the processed at least one periodic data set to obtain a feature map;
residual convolution unit construction unit 13: and constructing a residual convolution unit, and extracting spatial correlation characteristics according to the trend data set and the characteristic diagram.
The data set constructing unit 11 selects the cycle data set and the trend data set for the prediction target at different time periods according to the spatial correlation data.
Wherein the at least one periodic data set includes at least one of a weekly interval historical period data set, a daily interval historical period data set, a monthly interval historical period data set, a yearly interval historical period data set, and a current day trend data set, the trend data set being a current day trend data set.
Further, the feature map obtaining unit 12 includes:
the data set processing module 121: processing at least one periodic data set based on a time sliding mechanism to correspondingly obtain at least one sliding periodic data set;
the feature map obtaining module 122: and performing feature extraction and aggregation on at least one sliding period data set through a two-dimensional point-by-point convolution layer to obtain a feature map.
Still further, the residual convolution unit construction unit 13 includes:
the dependent feature extraction module 131: extracting the characteristics of departure place dependence and destination dependence by stacking a plurality of depth-separated one-dimensional convolution layers;
the spatial correlation feature output module 132: and extracting a spatial correlation structure through the one-dimensional point-by-point convolutional layer, and aggregating the departure place dependence and destination dependence characteristics to output spatial correlation characteristics.
In conclusion, for the complex spatial correlation of the spatial correlation data, a residual convolution unit is constructed and the characteristics of departure place dependence and destination dependence are identified, so that the accurate description of the complex spatial dependence is realized, and more accurate data support is provided for the future time-space prediction.
Although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.
Claims (10)
1. A method for extracting spatial correlation features is characterized by comprising the following steps:
step S1: selecting the spatial correlation data to obtain at least one periodic data set and a trend data set;
step S2: processing at least one periodic data set based on a time sliding mechanism, and performing feature extraction and aggregation on the processed at least one periodic data set to obtain a feature map;
step S3: and constructing a residual convolution unit, and extracting spatial correlation characteristics according to the trend data set and the characteristic diagram.
2. The extraction method according to claim 1, wherein the step S1 includes:
and selecting at least one periodic data set and at least one trend data set by taking different time periods as prediction targets according to the spatial correlation data.
3. The extraction method of any one of claims 1-2, wherein the at least one periodic data set includes at least one of a weekly interval historical period data set, a daily interval historical period data set, a monthly interval historical period data set, and an annual interval historical period data set, and the trend data set is a current day trend data set.
4. The extraction method according to claim 2, wherein the step S2 includes:
step S21: processing at least one periodic data set based on a time sliding mechanism to correspondingly obtain at least one sliding periodic data set;
step S22: and performing feature extraction and aggregation on at least one sliding period data set through a two-dimensional point-by-point convolution layer to obtain a feature map.
5. The extraction method according to claim 4, wherein the step S3 includes:
step S31: extracting the characteristics of departure place dependence and destination dependence by stacking a plurality of depth-separated one-dimensional convolution layers;
step S32: and extracting a spatial correlation structure through the one-dimensional point-by-point convolutional layer, and aggregating the departure place dependence and destination dependence characteristics to output spatial correlation characteristics.
6. A system for extracting spatially correlated features, comprising:
a data set construction unit: selecting the spatial correlation data to obtain at least one periodic data set and a trend data set;
a feature map obtaining unit: processing at least one periodic data set based on a time sliding mechanism, and performing feature extraction and aggregation on the processed at least one periodic data set to obtain a feature map;
residual convolution unit construction unit: and constructing a residual convolution unit, and extracting spatial correlation characteristics according to the trend data set and the characteristic diagram.
7. The extraction system according to claim 6, wherein the data set construction unit selects at least one periodic data set and a trend data set for the prediction target at different time periods based on the spatial correlation data.
8. The extraction system of any one of claims 6-7, wherein the at least one periodic data set includes at least one of a weekly interval historical period data set, a daily interval historical period data set, a monthly interval historical period data set, and an annual interval historical period data set, the trend data set being a current day trend data set.
9. The extraction system according to claim 8, wherein the feature map obtaining unit includes:
a data set processing module: processing at least one periodic data set based on a time sliding mechanism to correspondingly obtain at least one sliding periodic data set;
a characteristic diagram obtaining module: and performing feature extraction and aggregation on at least one sliding period data set through a two-dimensional point-by-point convolution layer to obtain a feature map.
10. The extraction system of claim 9, wherein said residual convolution unit construction unit comprises:
a dependent feature extraction module: extracting the characteristics of departure place dependence and destination dependence by stacking a plurality of depth-separated one-dimensional convolution layers;
the spatial correlation characteristic output module: and extracting a spatial correlation structure through the one-dimensional point-by-point convolutional layer, and aggregating the departure place dependence and destination dependence characteristics to output spatial correlation characteristics.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010336499.9A CN111242394B (en) | 2020-04-26 | 2020-04-26 | Method and system for extracting spatial correlation characteristics |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202010336499.9A CN111242394B (en) | 2020-04-26 | 2020-04-26 | Method and system for extracting spatial correlation characteristics |
Publications (2)
Publication Number | Publication Date |
---|---|
CN111242394A true CN111242394A (en) | 2020-06-05 |
CN111242394B CN111242394B (en) | 2021-03-23 |
Family
ID=70864691
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202010336499.9A Active CN111242394B (en) | 2020-04-26 | 2020-04-26 | Method and system for extracting spatial correlation characteristics |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN111242394B (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6134500A (en) * | 1999-06-03 | 2000-10-17 | United Air Lines, Inc. | System and method for generating optimal flight plans for airline operations control |
CN108985475A (en) * | 2018-06-13 | 2018-12-11 | 厦门大学 | Net based on deep neural network about vehicle car service needing forecasting method |
CN109492814A (en) * | 2018-11-15 | 2019-03-19 | 中国科学院深圳先进技术研究院 | A kind of Forecast of Urban Traffic Flow prediction technique, system and electronic equipment |
CN110782070A (en) * | 2019-09-25 | 2020-02-11 | 北京市交通信息中心 | Urban rail transit emergency passenger flow space-time distribution prediction method |
-
2020
- 2020-04-26 CN CN202010336499.9A patent/CN111242394B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US6134500A (en) * | 1999-06-03 | 2000-10-17 | United Air Lines, Inc. | System and method for generating optimal flight plans for airline operations control |
CN108985475A (en) * | 2018-06-13 | 2018-12-11 | 厦门大学 | Net based on deep neural network about vehicle car service needing forecasting method |
CN109492814A (en) * | 2018-11-15 | 2019-03-19 | 中国科学院深圳先进技术研究院 | A kind of Forecast of Urban Traffic Flow prediction technique, system and electronic equipment |
CN110782070A (en) * | 2019-09-25 | 2020-02-11 | 北京市交通信息中心 | Urban rail transit emergency passenger flow space-time distribution prediction method |
Non-Patent Citations (1)
Title |
---|
张国兴: "具有长期依赖的出租车轨迹目的地预测", 《中国优秀硕士学位论文全文数据库》 * |
Also Published As
Publication number | Publication date |
---|---|
CN111242394B (en) | 2021-03-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
Bi et al. | Daily tourism volume forecasting for tourist attractions | |
CN111242292B (en) | OD data prediction method and system based on deep space-time network | |
CN111242395B (en) | Method and device for constructing prediction model for OD (origin-destination) data | |
CN111223301B (en) | Traffic flow prediction method based on graph attention convolution network | |
Pan et al. | Predicting bike sharing demand using recurrent neural networks | |
CN110991713B (en) | Irregular area flow prediction method based on multi-graph convolution sum GRU | |
CN110766942B (en) | Traffic network congestion prediction method based on convolution long-term and short-term memory network | |
CN110503104B (en) | Short-time remaining parking space quantity prediction method based on convolutional neural network | |
US20240054321A1 (en) | Traffic prediction | |
CN112561187B (en) | Network taxi booking target order prediction method based on CNN-LSTM | |
Venkatesh et al. | Rainfall prediction using generative adversarial networks with convolution neural network | |
CN113762595B (en) | Traffic time prediction model training method, traffic time prediction method and equipment | |
CN115204478A (en) | Public traffic flow prediction method combining urban interest points and space-time causal relationship | |
CN115545758B (en) | Method and system for self-adaptive incremental site selection of urban service facilities | |
CN112488185A (en) | Method, system, electronic device and readable storage medium for predicting vehicle operating parameters including spatiotemporal characteristics | |
CN117829375B (en) | Method, device, equipment and medium for predicting multi-region demand of inter-city shuttle passenger transport | |
CN117494034A (en) | Air quality prediction method based on traffic congestion index and multi-source data fusion | |
CN115762147A (en) | Traffic flow prediction method based on adaptive graph attention neural network | |
CN116913088A (en) | Intelligent flow prediction method for expressway | |
CN111985731B (en) | Method and system for predicting number of people at urban public transport station | |
Wang et al. | MetroEye: A weather-aware system for real-time metro passenger flow prediction | |
CN111242394B (en) | Method and system for extracting spatial correlation characteristics | |
Ge et al. | Traffic speed prediction with missing data based on TGCN | |
Li et al. | How spatial features affect urban rail transit prediction accuracy: A deep learning based passenger flow prediction method | |
CN113344290B (en) | Method for correcting sub-season rainfall weather forecast based on U-Net network |
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 |