CN114692022A - Position prediction method and system based on space-time behavior mode - Google Patents

Position prediction method and system based on space-time behavior mode Download PDF

Info

Publication number
CN114692022A
CN114692022A CN202210346045.9A CN202210346045A CN114692022A CN 114692022 A CN114692022 A CN 114692022A CN 202210346045 A CN202210346045 A CN 202210346045A CN 114692022 A CN114692022 A CN 114692022A
Authority
CN
China
Prior art keywords
term
acquiring
long
result
user
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.)
Pending
Application number
CN202210346045.9A
Other languages
Chinese (zh)
Inventor
胡瑞敏
李希希
王正
王晓晨
黄文心
徐东曙
柴笑宇
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 CN202210346045.9A priority Critical patent/CN114692022A/en
Publication of CN114692022A publication Critical patent/CN114692022A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9537—Spatial or temporal dependent retrieval, e.g. spatiotemporal queries
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
    • G06F16/29—Geographical information databases
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • General Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • Biophysics (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Computational Linguistics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Health & Medical Sciences (AREA)
  • Remote Sensing (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

The invention discloses a position prediction method and a position prediction system based on a space-time behavior mode. The method comprises the following steps: step S1, acquiring an embedded expression vector according to the historical track data of the user; step S2, constructing a plurality of recurrent neural networks; s3, obtaining a splicing result of a part of the cyclic neural network, inputting the splicing result to a full connection layer, and obtaining long-term regularity of a user in a plurality of feature spaces; s4, acquiring the hidden layer state of the residual recurrent neural network and the output result of the attention mechanism, inputting the output result to the full connection layer, and acquiring the short-term randomness of the user time-space behavior pattern; and step S5, fusing the long-term regularity and the short-term randomness obtained by the processing to obtain a final position prediction result. According to the position prediction method based on the space-time behavior mode, the accuracy of position prediction is improved by considering long-term regularity and short-term randomness.

Description

Position prediction method and system based on space-time behavior mode
Technical Field
The invention relates to the technical field of information prediction, in particular to a position prediction method and a position prediction system based on a space-time behavior mode.
Background
Location prediction is a technique for predicting a future location where a user is likely to appear based on historical trajectory data of the user over a period of time in the past. The method is generally used in the fields of smart cities, interest point recommendation, crime prediction and the like, so the position prediction is a research hotspot of current big data analysis and multimedia application.
In recent years, there has been a lot of research work on position prediction. These research works can be largely divided into three major categories: matrix factorization based methods, markov chain based methods, deep network based methods. However, the spatiotemporal trajectory information of people has multiple properties, such as time, space and social space, and the records left on the social network by the users also contain a lot of semantic texts from which multiple features can be extracted. In addition, the behavior pattern of a person is not constant, not only regular, but also random disturbances. Therefore, position prediction remains a difficult task.
Disclosure of Invention
The invention aims to solve the problems of data multivariate isomerism and behavior randomness in the existing similar algorithm and provides a position prediction method and a position prediction system based on a space-time behavior mode.
In a first aspect, the present application provides a position prediction method based on a spatiotemporal behavior pattern, including the following steps:
step S1, acquiring an embedded expression vector according to the historical track data of the user;
step S2, constructing a plurality of recurrent neural networks;
s3, obtaining a splicing result of a part of the cyclic neural network, inputting the splicing result to a full connection layer, and obtaining long-term regularity of a user in a plurality of feature spaces;
s4, acquiring the hidden layer state of the residual recurrent neural network and the output result of the attention mechanism, inputting the output result to the full connection layer, and acquiring the short-term randomness of the user time-space behavior pattern;
and step S5, fusing the long-term regularity and the short-term randomness obtained by the processing to obtain a final position prediction result.
According to the first aspect, in a first possible implementation manner of the first aspect, the step S1 includes the following steps:
step S11, acquiring historical track data of a user;
step S12, acquiring sequence information extracted from historical track data, wherein the sequence information comprises a time sequence, a date sequence, a place sequence, an activity type sequence and a text vector sequence;
and step S13, inputting the sequence information into the embedding layer to obtain the embedding expression vector.
According to the first aspect, in a second possible implementation manner of the first aspect, the step S2 specifically includes the following steps:
and step S20, constructing a plurality of recurrent neural networks, wherein the network length of part of recurrent neural networks is M, the length of the rest recurrent neural networks is N, and M and N are unequal.
According to the second possible implementation manner of the first aspect, in a third possible implementation manner of the first aspect, after the step S20, the method further includes the following steps:
and constructing a loss function model by combining the tags, and obtaining the optimized recurrent neural network through optimization training of an ADAM (adaptive dynamic analysis) optimization algorithm.
According to the second possible implementation manner of the first aspect, in a fourth possible implementation manner of the first aspect, the step S3 specifically includes the following steps:
s31, obtaining a splicing result of the output result of the last time of the partial cyclic neural network;
and step S32, inputting the splicing result to a full connection layer, and acquiring the long-term regularity of the user in a plurality of feature spaces.
According to the fourth possible implementation manner of the first aspect, in a fifth possible implementation manner of the first aspect, the step S4 includes the following steps:
step S41, adding an attention mechanism to adjust the hidden layer state value of the residual neural network;
step S42, obtaining an output result according to the adjusted hidden layer state value;
and step S43, inputting and outputting the result to a full connection layer, and acquiring the short-term randomness of the user space-time behavior pattern.
According to the fifth possible implementation manner of the first aspect, in a sixth possible implementation manner of the first aspect, the step S5 specifically includes the following steps:
and fusing the long-term regularity and the short-term randomness obtained by processing to obtain a final position prediction result according to the following formula:
Pf＝γ*Plong+(1-γ)*Pshort，
wherein, PlongFor long-term regularity, PshortIs short-term randomness.
In a second aspect, the present application provides a position prediction system based on spatiotemporal behavior patterns, comprising:
the embedded expression vector acquisition module is used for acquiring an embedded expression vector according to historical track data of a user;
the neural network construction module is used for constructing a plurality of cyclic neural networks;
the long-term regularity acquisition module is in communication connection with the embedded expression vector acquisition module and the neural network construction module and is used for acquiring the splicing result of a part of circulating neural networks, inputting the splicing result to a full connection layer and acquiring the long-term regularity of the user in a plurality of feature spaces;
the short-term randomness acquisition module is in communication connection with the embedded expression vector acquisition module and the neural network construction module, and is used for acquiring the hidden layer state of the residual cyclic neural network and the output result of the attention mechanism, inputting the output result to the full connection layer, and acquiring the short-term randomness of the user time-space behavior pattern;
and the final position result acquisition module is in communication connection with the long-term regularity acquisition module and the short-term randomness acquisition module and is used for fusing the acquired long-term regularity and short-term randomness to acquire a final position prediction result.
According to the second aspect, in a first possible implementation manner of the second aspect, the embedded expression vector obtaining module includes:
a historical track data acquisition unit for acquiring historical track data of a user;
the sequence information acquisition unit is in communication connection with the historical track data acquisition unit and is used for acquiring sequence information extracted from the historical track data, and the sequence information comprises a time sequence, a date sequence, a place sequence, an activity type sequence and a text vector sequence;
and the embedded expression vector acquisition unit is in communication connection with the sequence information acquisition unit and is used for inputting sequence information into an embedded layer and acquiring an embedded expression vector.
According to the second aspect, in a second possible implementation manner of the second aspect, the long-term regularity obtaining module includes:
the splicing result acquisition unit is used for acquiring the splicing result of the output result of the last time of the partial cyclic neural network;
and the long-term regularity acquisition unit is in communication connection with the splicing result acquisition unit and is used for inputting the splicing result to the full connection layer and acquiring the long-term regularity of the user in a plurality of feature spaces.
Compared with the prior art, the invention has the following advantages:
according to the position prediction method based on the space-time behavior mode, the accuracy of position prediction is improved by considering long-term regularity and short-term randomness.
Drawings
FIG. 1 is a flow chart of a method of a spatiotemporal behavior pattern-based location prediction method according to an embodiment of the present invention;
FIG. 2 is a flow chart of another method of a spatiotemporal behavior pattern-based location prediction method according to an embodiment of the present invention;
FIG. 3 is a functional block diagram of a spatiotemporal behavior pattern-based location prediction system according to an embodiment of the present invention;
FIG. 4 is another functional block diagram of a spatiotemporal behavior pattern-based location prediction system according to an embodiment of the present invention.
Detailed Description
Reference will now be made in detail to the present embodiments of the invention, examples of which are illustrated in the accompanying drawings. While the invention will be described in conjunction with the specific embodiments, it will be understood that they are not intended to limit the invention to the embodiments described. On the contrary, it is intended to cover alternatives, modifications, and equivalents as may be included within the spirit and scope of the invention as defined by the appended claims. It should be noted that the method steps described herein may be implemented by any functional block or functional arrangement, and that any functional block or functional arrangement may be implemented as a physical entity or a logical entity, or a combination of both.
In order that those skilled in the art will better understand the present invention, the following detailed description of the invention is provided in conjunction with the accompanying drawings and the detailed description of the invention.
Note that: the example to be described next is only a specific example, and does not limit the embodiments of the present invention necessarily to the following specific steps, values, conditions, data, orders, and the like. Those skilled in the art can, upon reading this specification, utilize the concepts of the present invention to construct more embodiments than those specifically described herein.
The position prediction is a research hotspot of current big data analysis and multimedia application, the existing position prediction research work is mainly divided into three categories, namely a matrix decomposition-based method, a Markov chain-based method and a deep network-based method, but the existing similar algorithms have the problems of data multi-element isomerism and behavior randomness, so that the position prediction precision is not high.
Referring to fig. 1, an embodiment of the present invention provides a method and a system for position prediction based on a spatiotemporal behavior pattern, including the following steps:
step S1, acquiring an embedded expression vector according to the historical track data of the user;
step S2, constructing a plurality of recurrent neural networks;
s3, obtaining a splicing result of a part of the cyclic neural network, inputting the splicing result to a full connection layer, and obtaining long-term regularity of a user in a plurality of feature spaces;
s4, acquiring the hidden layer state of the residual recurrent neural network and the output result of the attention mechanism, inputting the output result to the full connection layer, and acquiring the short-term randomness of the user time-space behavior pattern;
and step S5, fusing the long-term regularity and the short-term randomness obtained by the processing to obtain a final position prediction result.
According to the position prediction method based on the space-time behavior mode, the accuracy of position prediction is improved by considering long-term regularity and short-term randomness.
In an embodiment, referring to fig. 2, the step S1 includes the following steps:
step S11, acquiring historical track data of a user, namely a spatiotemporal behavior pattern data set of the user, wherein the data set comprises information such as user ID, time information, date information, place ID, longitude and latitude coordinates of a place, activity type and the like;
step 1, the spatiotemporal behavior pattern data set of the user comprises: user ID, time information, date information, location ID, longitude and latitude coordinates of the location, activity type, and the like.
Acquiring historical track data of users, and expressing the historical track data of each user u as
Figure BDA0003576558810000061
Wherein, therein
Figure BDA0003576558810000062
Representing a single recording at time t,
Figure BDA0003576558810000063
Figure BDA0003576558810000064
the corresponding elements respectively represent user information, longitude and latitude coordinates of a place, date, hour, activity type, place, and time stamp.
And step S12, acquiring sequence information extracted from the historical track data, and acquiring track information of the multi-feature space of the user, wherein the sequence information comprises a time sequence, a date sequence, a place sequence, an activity type sequence and a text vector sequence.
Extracting from t in historical track data1Time to tnTime of day, date, location, activity type and text vector, respectively denoted as g, d, l, c, x. Specifically, each time instant is composed of different features to form a text vector x ═ d, g, c, l, ttr,geotr},ttr,geotrRespectively representing the time interval and the geographical distance between two check-in points, from t1Time to tnThe sequence of time instants is denoted x.
And step S13, inputting the sequence information into the embedding layer to obtain the embedding expression vector.
In a more specific embodiment, the inputs g, d, l, c, x, respectively, have corresponding outputs denoted as eg，ed，el，ec，ex。
In an embodiment, the step S2 specifically includes the following steps:
and step S20, constructing a plurality of recurrent neural networks, wherein the network length of part of the recurrent neural networks is M, the length of the rest recurrent neural networks is N, and the M and N are unequal.
In a more specific embodiment, 5 recurrent neural networks are constructed, respectively labeled as network1, network2, network3, network4, and network5, wherein the length of network5 is M, the length of network1, the length of network2, the length of network3, and the length of network4 are all N.
In an embodiment, after the step S20, the method further includes the following steps:
and constructing a loss function model by combining the tags, and obtaining the optimized recurrent neural network through optimization training of an ADAM (adaptive dynamic adaptive analysis) optimization algorithm.
According to the second possible implementation manner of the first aspect, in a fourth possible implementation manner of the first aspect, the step S3 specifically includes the following steps:
step S31, obtaining a splicing result of the output result of the last time of the partial cyclic neural network;
and step S32, inputting the splicing result to a full connection layer, and acquiring the long-term regularity of the user in a plurality of feature spaces.
In an embodiment, before the step S31, the method further includes the following steps:
embedding expression vector embedding expression eg，ed，el，ec，exRespectively, to network1, network2, network3, network4, and network 5. Wherein, the hidden layer iteration process of the network1 is as follows:
Figure BDA0003576558810000081
Figure BDA0003576558810000082
Figure BDA0003576558810000083
Figure BDA0003576558810000084
Figure BDA0003576558810000085
wherein at tiAt the moment, input as
Figure BDA0003576558810000086
The hidden layer state is
Figure BDA0003576558810000087
Output is as
Figure BDA0003576558810000088
r and z denote the reset gate and update gate, respectively, of the recurrent neural network GRU at tiThe state value at the moment, sigma, represents a sigmod function, and tanh represents an activation function; w is a group ofr，Wz，W，WoIs a parameter to be learned; the hidden layer at the last moment is taken as the final output result of the network module. The network structures of network1, network2, network3 and network4 are the same, the difference is that the input and output are different, and the output result is respectively represented as PG，PD，PL，PC。
In a more specific embodiment, the step S31 is specifically implemented as:
the last time t of network1, network2, network3 and network4NThe output results are spliced into a whole, namely the long-term behavior rule P of the userlong＝[PG,PD,PL,PC]。
In one embodiment, the step S4 includes the following steps:
step S41, adding an attention mechanism to adjust the hidden layer state value of the residual neural network;
step S42, obtaining an output result according to the adjusted hidden layer state value;
and step S43, inputting the output result of the step S42 to a full connection layer, and acquiring the short-term randomness of the user space-time behavior pattern.
As mentioned above, the fully-connected layer is a full connected layer, the core operation of the fully-connected layer in the neural network is a matrix-vector product, i.e. y ═ Wx, and the role of the fully-connected layer is mainly to linearly transform the feature space to another feature space, which is equivalent to the role of the classifier.
In an embodiment, the step S4 is specifically implemented as:
adding a hidden layer of network5 into an attention mechanism, wherein the hidden layer iteration process of network5 is the same as that of network1, the difference is that the lengths of network modules are different, and the attention mechanism is added to the network5 to adjust the state value of the hidden layer of the recurrent neural network of the network5
Figure BDA0003576558810000091
The specific formula of the weight calculation of the attention mechanism is as follows:
Figure BDA0003576558810000092
Figure BDA0003576558810000093
x(ti)’＝α(ti)T·x(ti)；
Figure BDA0003576558810000094
it should be noted that
Figure BDA0003576558810000095
Is obtained by
Figure BDA0003576558810000096
And (4) calculating. Wherein, tanh represents the activation function,
Figure BDA0003576558810000097
Wshort，Ws1，Ws2，Ws3，bs1，bs2is a parameter to be trained, dxA dimension representing a text vector x is shown,
Figure BDA0003576558810000098
representing weights corresponding to hidden layer states after attention mechanism addition, e.g. the dimension of data feature x is dxThen x(ti)’＝α(ti)T·x(ti) Representing that each dimension of the feature value of the feature x is respectively assigned with a weight
Figure BDA0003576558810000099
Thus, a new feature vector is obtained, T being the time or period. x (t)i) ' indicates a text vector updated after the attention mechanism is added,
Figure BDA00035765588100000910
represents tiAnd (4) outputting the time. The last output is the last moment t after the attention mechanism is introduced for updatingMIs marked as Pshort。
In an embodiment, the step S5 specifically includes the following steps:
inputting the obtained long-term regularity and short-term randomness into a full-link layer and fusing the obtained long-term regularity and short-term randomness for position prediction to obtain a final position prediction result:
Pf＝γ*Plong+(1-γ)*Pshort，
wherein, PlongFor long-term regularity, PshortFor short-term randomness, gamma is a numerical value obtained by training, and gamma is more than 0 and less than 1.
Based on the same inventive concept, please refer to fig. 3, the present application provides a position prediction system based on spatiotemporal behavior pattern, comprising:
an embedded expression vector acquisition module 100 for acquiring an embedded expression vector according to the historical track data of the user;
a neural network construction module 200, configured to construct a plurality of recurrent neural networks;
a long-term regularity obtaining module 300, communicatively connected to the embedded expression vector obtaining module 100 and the neural network constructing module 200, configured to obtain a splicing result of a partial cyclic neural network, input the splicing result to a full connection layer, and obtain long-term regularity of a user in a plurality of feature spaces;
a short-term randomness obtaining module 400, communicatively connected to the embedded expression vector obtaining module 100 and the neural network constructing module 200, for obtaining the hidden layer state of the remaining recurrent neural network and the output result of the attention mechanism, inputting the output result to the full connection layer, and obtaining the short-term randomness of the user time-space behavior pattern;
and a final position result obtaining module, communicatively connected to the long-term regularity obtaining module 300 and the short-term randomness obtaining module 400, for merging the obtained long-term regularity and short-term randomness to obtain a final position prediction result.
In an embodiment, referring to fig. 4, the embedded expression vector obtaining module 100 includes:
a historical track data acquisition unit 110 configured to acquire historical track data of a user;
a sequence information obtaining unit 120, communicatively connected to the historical track data obtaining unit 110, configured to obtain sequence information extracted from the historical track data, where the sequence information includes a time sequence, a date sequence, a place sequence, an activity type sequence, and a text vector sequence;
an embedded expression vector obtaining unit 130, communicatively connected to the sequence information obtaining unit 120, for inputting sequence information into an embedded layer to obtain an embedded expression vector.
In one embodiment, the long-term regularity obtaining module includes:
the splicing result acquisition unit is used for acquiring the splicing result of the output result of the last time of the partial cyclic neural network;
and the long-term regularity acquisition unit is in communication connection with the splicing result acquisition unit and is used for inputting the splicing result to the full connection layer and acquiring the long-term regularity of the user in a plurality of feature spaces.
Based on the same inventive concept, the embodiments of the present application further provide a computer-readable storage medium, on which a computer program is stored, and the computer program, when executed by a processor, implements all or part of the method steps of the above method.
The present invention can implement all or part of the processes of the above methods, and can also be implemented by using a computer program to instruct related hardware, where the computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of the above method embodiments can be implemented. Wherein the computer program comprises computer program code, which may be in the form of source code, object code, an executable file or some intermediate form, etc. The computer readable medium may include: any entity or device capable of carrying computer program code, recording medium, U.S. disk, removable hard disk, magnetic disk, optical disk, computer Memory, Read-Only Memory (ROM), Random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution media, and the like. It should be noted that the computer-readable medium may contain suitable additions or subtractions depending on the requirements of legislation and patent practice in jurisdictions, for example, in some jurisdictions, computer-readable media may not include electrical carrier signals or telecommunication signals in accordance with legislation and patent practice.
Based on the same inventive concept, an embodiment of the present application further provides an electronic device, which includes a memory and a processor, where the memory stores a computer program running on the processor, and the processor executes the computer program to implement all or part of the method steps in the method.
The Processor may be a Central Processing Unit (CPU), other general purpose Processor, a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), an off-the-shelf Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic, discrete hardware components, etc. The general purpose processor may be a microprocessor or the processor may be any conventional processor or the like, the processor being the control center of the computer device and the various interfaces and lines connecting the various parts of the overall computer device.
The memory may be used to store computer programs and/or modules, and the processor may implement various functions of the computer device by executing or executing the computer programs and/or modules stored in the memory, as well as by invoking data stored in the memory. The memory may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required by at least one function (e.g., a sound playing function, an image playing function, etc.); the storage data area may store data (e.g., audio data, video data, etc.) created according to the use of the cellular phone. In addition, the memory may include high speed random access memory, and may also include non-volatile memory, such as a hard disk, a memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) Card, a Flash memory Card (Flash Card), at least one magnetic disk storage device, a Flash memory device, or other volatile solid state storage device.
As will be appreciated by one skilled in the art, embodiments of the present invention may be provided as a method, system, server, or computer program product. Accordingly, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present invention may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage, and the like) having computer-usable program code embodied therein.
The present invention has been described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), servers and computer program products according to embodiments of the invention. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
It will be apparent to those skilled in the art that various changes and modifications may be made in the present invention without departing from the spirit and scope of the invention. Thus, if such modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include such modifications and variations.

Claims (10)

1. A position prediction method based on a space-time behavior mode is characterized by comprising the following steps:
step S1, acquiring an embedded expression vector according to the historical track data of the user;
step S2, constructing a plurality of recurrent neural networks;
s3, obtaining a splicing result of a part of the cyclic neural network, inputting the splicing result to a full connection layer, and obtaining long-term regularity of a user in a plurality of feature spaces;
s4, acquiring the hidden layer state of the residual recurrent neural network and the output result of the attention mechanism, inputting the output result to the full connection layer, and acquiring the short-term randomness of the user time-space behavior pattern;
and step S5, fusing the long-term regularity and the short-term randomness obtained by the processing to obtain a final position prediction result.
2. The spatio-temporal behavior pattern-based location prediction method according to claim 1, wherein the step S1 comprises the steps of:
step S11, acquiring historical track data of a user;
step S12, acquiring sequence information extracted from historical track data, wherein the sequence information comprises a time sequence, a date sequence, a place sequence, an activity type sequence and a text vector sequence;
and step S13, inputting the sequence information into the embedding layer to obtain the embedding expression vector.
3. The spatio-temporal behavior pattern-based location prediction method according to claim 1, wherein the step S2 specifically comprises the steps of:
and step S20, constructing a plurality of recurrent neural networks, wherein the network length of part of recurrent neural networks is M, the length of the rest recurrent neural networks is N, and M and N are unequal.
4. The spatiotemporal behavior pattern-based location prediction method as defined in claim 3, further comprising, after the step S20, the steps of:
and constructing a loss function model by combining the tags, and obtaining the optimized recurrent neural network through optimization training of an ADAM (adaptive dynamic adaptive analysis) optimization algorithm.
5. The spatio-temporal behavior pattern-based location prediction method according to claim 3, wherein the step S3 specifically comprises the steps of:
s31, obtaining a splicing result of the output result of the last time of the partial cyclic neural network;
and step S32, inputting the splicing result to a full connection layer, and acquiring the long-term regularity of the user in a plurality of feature spaces.
6. The spatio-temporal behavior pattern-based location prediction method according to claim 5, wherein the step S4 comprises the steps of:
step S41, adding an attention mechanism to adjust the hidden layer state value of the residual neural network;
step S42, obtaining an output result according to the adjusted hidden layer state value;
and step S43, inputting and outputting the result to a full connection layer, and acquiring the short-term randomness of the user space-time behavior pattern.
7. The method for spatio-temporal behavior pattern-based location prediction as defined in claim 6, wherein said step S5 specifically comprises the steps of:
and fusing the long-term regularity and the short-term randomness obtained by processing to obtain a final position prediction result according to the following formula:
Pf＝γ*Plong+(1-γ)*Pshort，
wherein, PlongFor long-term regularity, PshortIs short-term randomness.
8. A system for location prediction based on spatiotemporal behavioral patterns, comprising:
the embedded expression vector acquisition module is used for acquiring an embedded expression vector according to historical track data of a user;
the neural network construction module is used for constructing a plurality of cyclic neural networks;
the long-term regularity acquisition module is in communication connection with the embedded expression vector acquisition module and the neural network construction module and is used for acquiring the splicing result of a part of circulating neural networks, inputting the splicing result to a full connection layer and acquiring the long-term regularity of the user in a plurality of feature spaces;
the short-term randomness acquisition module is in communication connection with the embedded expression vector acquisition module and the neural network construction module and is used for acquiring the hidden layer state of the residual cyclic neural network and the output result of the attention mechanism, inputting the output result to the full connection layer and acquiring the short-term randomness of the user time-space behavior pattern;
and the final position result acquisition module is in communication connection with the long-term regularity acquisition module and the short-term randomness acquisition module and is used for fusing the acquired long-term regularity and short-term randomness to acquire a final position prediction result.
9. The spatiotemporal behavior pattern-based position prediction system of claim 8, wherein the embedded expression vector acquisition module comprises:
a historical track data acquisition unit for acquiring historical track data of a user;
the sequence information acquisition unit is in communication connection with the historical track data acquisition unit and is used for acquiring sequence information extracted from the historical track data, and the sequence information comprises a time sequence, a date sequence, a place sequence, an activity type sequence and a text vector sequence;
and the embedded expression vector acquisition unit is in communication connection with the sequence information acquisition unit and is used for inputting sequence information into an embedded layer and acquiring an embedded expression vector.
10. The spatiotemporal behavior pattern-based location prediction system of claim 8, wherein the long-term regularity acquisition module comprises:
the splicing result acquisition unit is used for acquiring the splicing result of the output result of the last time of the partial cyclic neural network;
and the long-term regularity acquisition unit is in communication connection with the splicing result acquisition unit and is used for inputting the splicing result to the full connection layer and acquiring the long-term regularity of the user in a plurality of feature spaces.
CN202210346045.9A 2022-03-31 2022-03-31 Position prediction method and system based on space-time behavior mode Pending CN114692022A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210346045.9A CN114692022A (en) 2022-03-31 2022-03-31 Position prediction method and system based on space-time behavior mode

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210346045.9A CN114692022A (en) 2022-03-31 2022-03-31 Position prediction method and system based on space-time behavior mode

Publications (1)

Publication Number Publication Date
CN114692022A true CN114692022A (en) 2022-07-01

Family

ID=82140603

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210346045.9A Pending CN114692022A (en) 2022-03-31 2022-03-31 Position prediction method and system based on space-time behavior mode

Country Status (1)

Country Link
CN (1) CN114692022A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117232097A (en) * 2023-11-09 2023-12-15 上海轻环能源科技有限公司 Central air conditioner refrigerating station optimal control method and system based on self-learning fusion model

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117232097A (en) * 2023-11-09 2023-12-15 上海轻环能源科技有限公司 Central air conditioner refrigerating station optimal control method and system based on self-learning fusion model
CN117232097B (en) * 2023-11-09 2024-02-20 上海轻环能源科技有限公司 Central air conditioner refrigerating station optimal control method and system based on self-learning fusion model

Similar Documents

Publication Publication Date Title
CN109034861B (en) User loss prediction method and device based on mobile terminal log behavior data
CN111461168B (en) Training sample expansion method and device, electronic equipment and storage medium
US11423325B2 (en) Regression for metric dataset
CN110728317A (en) Training method and system of decision tree model, storage medium and prediction method
CN110659678B (en) User behavior classification method, system and storage medium
Xu et al. Bayesian non-parametric inference for stochastic epidemic models using Gaussian processes
CN109543112A (en) A kind of sequence of recommendation method and device based on cyclic convolution neural network
CN112988851B (en) Counterfactual prediction model data processing method, device, equipment and storage medium
CN112085541A (en) User demand analysis method and device based on browsing consumption time series data
CN106909560A (en) Point of interest sort method
CN113656699A (en) User feature vector determination method, related device and medium
CN112966818A (en) Directional guide model pruning method, system, equipment and storage medium
US20210239479A1 (en) Predicted Destination by User Behavior Learning
CN114692022A (en) Position prediction method and system based on space-time behavior mode
CN113326877A (en) Model training method, data processing method, device, apparatus, storage medium, and program
US20180285758A1 (en) Methods for creating and analyzing dynamic trail networks
CN116542328B (en) Knowledge distillation method and device for CTR prediction model
CN110717537B (en) Method and device for training user classification model and executing user classification prediction
CN117056595A (en) Interactive project recommendation method and device and computer readable storage medium
Sunitha et al. Political optimizer-based automated machine learning for skin lesion data
CN114267422B (en) Method and system for predicting surface water quality parameters, computer equipment and storage medium
CN111694814B (en) Batch expansion method and device for date partition table, computer equipment and storage medium
CN114693352A (en) Client access amount prediction method and device
CN113035286A (en) Method for generating new chemical structure, neural network device, and non-transitory computer-readable medium
CN113192315A (en) Traffic flow distribution prediction method, prediction device and terminal equipment

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