CN110019616B - POI (Point of interest) situation acquisition method and equipment, storage medium and server thereof - Google Patents

POI (Point of interest) situation acquisition method and equipment, storage medium and server thereof Download PDF

Info

Publication number
CN110019616B
CN110019616B CN201711260201.5A CN201711260201A CN110019616B CN 110019616 B CN110019616 B CN 110019616B CN 201711260201 A CN201711260201 A CN 201711260201A CN 110019616 B CN110019616 B CN 110019616B
Authority
CN
China
Prior art keywords
poi
target poi
situation
pois
target
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN201711260201.5A
Other languages
Chinese (zh)
Other versions
CN110019616A (en
Inventor
杜逸康
龚剑
陈永全
泮华杰
成睿
张金晶
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.)
Tencent Technology Shenzhen Co Ltd
Original Assignee
Tencent Technology Shenzhen Co Ltd
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 Tencent Technology Shenzhen Co Ltd filed Critical Tencent Technology Shenzhen Co Ltd
Priority to CN201711260201.5A priority Critical patent/CN110019616B/en
Publication of CN110019616A publication Critical patent/CN110019616A/en
Application granted granted Critical
Publication of CN110019616B publication Critical patent/CN110019616B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • 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
    • 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
    • G06F16/9535—Search customisation based on user profiles and personalisation
    • 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

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Remote Sensing (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

The embodiment of the invention discloses a method for acquiring the situation state of a POI (point of interest), which comprises the following steps: acquiring network information characteristic data associated with a target POI; converting the acquired network information characteristic data into corresponding present characteristic vectors; inputting the converted situation characteristic vector of the target POI into a situation state judgment model to obtain the situation state of the target POI; and outputting the current state of the target POI. By adopting the method and the device, the POI situation in the electronic map data can be effectively ensured.

Description

POI (Point of interest) situation acquisition method and equipment, storage medium and server thereof
Technical Field
The invention relates to the technical field of internet, in particular to a method for acquiring the current state of a POI (point of interest), equipment, a storage medium and a server thereof.
Background
An electronic map, i.e., a digital map, is a map that is stored and referred to digitally using computer technology, and displays corresponding points of Interest (POI) or information points to a user according to different positions in the electronic map. The map presence refers to that the geographic spatial information provided by the map reflects the latest current situation as much as possible, and under the new situation of rapid city development, the POI presence changes very frequently, so that the POI presence and accuracy are crucial to electronic map services.
The method for updating the POI present state mainly comprises manual collection/reporting, needs a large amount of manpower and material resources, and has unsatisfactory updating efficiency and timeliness.
Disclosure of Invention
In view of this, embodiments of the present invention provide a method for acquiring a POI present status, which can effectively ensure the POI present status in electronic map data.
In order to solve the above technical problem, an embodiment of the present invention provides a method for acquiring a POI presence state, where the method includes:
acquiring network information characteristic data associated with a target POI;
converting the acquired network information characteristic data into corresponding present characteristic vectors;
inputting the converted current feature vector of the target POI into a current state judgment model to obtain the current state of the target POI;
and outputting the current state of the target POI.
Correspondingly, the embodiment of the invention also provides equipment for acquiring the POI present situation state, which comprises:
the network information acquisition module is used for acquiring network information characteristic data associated with a target POI;
the situation characteristic conversion module is used for converting the acquired network information characteristic data into corresponding situation characteristic vectors;
the situation state judgment module is used for inputting the situation characteristic vector of the target POI obtained by conversion into the situation state judgment model to obtain the situation state of the target POI;
and the current state output module is used for outputting the current state of the target POI.
Accordingly, embodiments of the present invention also provide a computer storage medium storing a plurality of instructions, the instructions being adapted to be loaded by a processor and to perform the following steps:
acquiring network information characteristic data associated with a target POI;
converting the acquired network information characteristic data into corresponding present characteristic vectors;
inputting the converted current feature vector of the target POI into a current state judgment model to obtain the current state of the target POI;
and outputting the current state of the target POI.
Correspondingly, an embodiment of the present invention further provides a server, including: a processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and to perform the steps of:
acquiring network information characteristic data associated with a target POI;
converting the acquired network information characteristic data into corresponding present characteristic vectors;
inputting the converted situation characteristic vector of the target POI into a situation state judgment model to obtain the situation state of the target POI;
and outputting the current state of the target POI.
According to the embodiment of the invention, the situation state of the target POI is rapidly judged according to the network information characteristic data and the situation state judgment model associated with the target POI, and inefficient manual data acquisition, reporting and auditing processes are not needed, so that the POI situation in the electronic map data is effectively ensured.
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, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic view of an implementation scenario architecture of a POI present status acquisition method according to an embodiment of the present invention;
fig. 2 is a schematic flowchart of a method for acquiring a POI present status according to an embodiment of the present invention;
FIG. 3 is a diagram illustrating a relationship between a presence status determination model and at least a presence status determination sub-model according to an embodiment of the present invention;
fig. 4 is a schematic diagram illustrating a current state of a target POI displayed on an electronic map according to an embodiment of the present invention;
FIG. 5 is a diagram illustrating a query result of a POI search request submitted in a search box according to an embodiment of the present invention is changed due to the POI presence status;
fig. 6 is a schematic flowchart of a POI presence state acquisition method according to another embodiment of the present invention;
FIG. 7 is a schematic diagram of a logical framework for implementing another embodiment of the present invention;
fig. 8 is a schematic flowchart of a POI presence state acquisition method according to yet another embodiment of the present invention;
fig. 9 is a schematic structural diagram of a POI present status acquiring apparatus in the embodiment of the present invention;
fig. 10 is a schematic structural diagram of a server according to an embodiment of the present invention.
Detailed Description
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 only a part of the embodiments of the present invention, and not all of the embodiments. 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.
The method for acquiring the POI present state can be applied to marking or updating the POI present state in the electronic map data, so that the electronic map data acquired by a user is more accurate and credible when the user acquires the geographic information service through the electronic map.
The POI present state acquisition method related to the embodiment of the invention is executed by depending on a computer program, and the POI present state acquisition device can be operated on a computer system of a Von Ruehmann system based on the POI present state. The POI present status acquiring Device may include a server having functions of storage, calculation, test, and the like, and may further include server devices such as a tablet computer, a Personal Computer (PC), a smart phone, a palmtop computer, and a Mobile Internet Device (MID).
The following describes the road condition information obtaining method according to the embodiment of the present invention in detail with reference to fig. 1 to 8.
Referring to fig. 1, which is a schematic view of an implementation scenario architecture of the POI present status acquiring method according to the embodiment of the present invention, a POI present status acquiring device may acquire POI-associated network information feature data from a plurality of network information feature data providers, and then process the POI-associated network information feature data to obtain a present status of a POI, on one hand, the present status of the POI may be output to an electronic map database, the electronic map database provides electronic map data to a user, and the electronic map database may further provide user behaviors generated by the user in a process of using the electronic map to the POI present status acquiring device; on the other hand, the POI present state acquisition equipment can maintain electronic map data per se, provide the electronic map data for the user per se and acquire user behaviors generated by the user in the process of using the electronic map.
Referring to fig. 2, a schematic flow chart of a method for acquiring a POI presence status is provided in an embodiment of the present invention. As shown in fig. 2, the method of the embodiment of the present invention may include the following steps S101 to S104.
S101, network information characteristic data associated with the target POI are obtained.
In the embodiment of the present invention, the target POI may be any POI or a part of POIs or all POIs in the electronic map data, and the network information feature data associated with the target POI may include access data of a wireless access point corresponding to the target POI, network user behavior data associated with the target POI, or network media data associated with the target POI.
The access data of the wireless access point corresponding to the target POI may be the number of access users of the wireless access point, which has previously established an association relationship with the target POI, within a period of time, for example, the total number of access users within the last several days or the number of access users per day within the last several days, where the several days may be 2 to 15 days. In specific implementation, the POI presence state acquiring device may collect access data of a plurality of wireless access points within a period of time, and then obtain access data of wireless access points corresponding to each POI in accordance with a pre-obtained association relationship between the plurality of wireless access points and the POI, where the association relationship between the wireless access points and the POI may be reported by a merchant corresponding to the POI, or may be provided by a manual acquisition or a third-party data provider, and a certain POI may correspond to a plurality of wireless access points, for example, one or more employee access points, one or more customer access points, or a background office access point. The access data of the wireless access point corresponding to the target POI can reflect the people flow condition of the target POI to a certain extent, so that the current state of the target POI can be judged.
The network user behavior data associated with the target POI may include data obtained by recording various network user behaviors performed by the user for the target POI within a period of time, where the various network user behaviors include click behaviors, navigation behaviors, payment behaviors, booking behaviors, comment behaviors, sharing behaviors, error reporting behaviors, and the like, for example, a click or selection operation, a viewing operation, a navigation operation, and the like input by the user for the target POI after retrieving the POI on an electronic map, and such network user behaviors performed for the target POI may reflect a degree of attention for the POI to some extent, so that a current state of the target POI may be determined. The user behavior recorded over time may be statistically obtained over the last several days, which may be, for example, 2-15 days.
The network media data can be various media data appearing in the network, such as information webpage data in webpage media, information data of news APP, articles, dynamics, messages and other data of social network media, the POI present state acquisition equipment can acquire a large amount of network media data from various media channels, then the network media data related to the POI are screened out, media segments related to the POI present state are further analyzed to obtain the media segments related to the POI present state, the media segments are used for judging the present state of the POI, for example, a news content is renovated for a certain building with a certain level of prosperous guests, then a target POI related to the news data is a necessary winner with an address located at one level of the building, and the news data can obviously serve as the network media data related to the target POI.
In an optional embodiment, the POI present status acquiring device may further use, according to an attachment relationship between the target POI and another POI, network information feature data of the another POI having an attachment relationship with the target POI as network information feature data associated with the target POI, that is, the network information feature data associated with the target POI may further include the network information feature data related to the target POI and the network information feature data related to the another POI having an attachment relationship with the target POI. The method comprises the steps that network information characteristic data related to a target POI, namely access data of a wireless access point corresponding to the target POI, network user behavior data related to the target POI or network media data related to the target POI are described as a first POI, and network information characteristic data related to other POIs which are in an attachment relationship with the target POI, namely access data of a wireless access point corresponding to the first POI, network user behavior data related to the first POI or network media data related to the first POI. For example, if restaurant a is located in building B or restaurant C is located in park D, the two POIs have an attached relationship with each other, and when the present state of the former POI is determined, the network information feature data of the latter POI may be taken into consideration as a reference. And if the POI with the unique dependency relationship exists at the same time, for example, only one F-type home supermarket exists in the E building, the network information characteristic data of any one of the two POIs of the E building and the F-type supermarket can be used as a reference for judging the current state of the other POI.
The attachment relationship between the POIs can be obtained by manually collecting or collecting POI address information and analyzing, natural language processing can be performed according to media text fragments containing the target POI name and other POI names, the attachment relationship between the target POI and other POIs can be obtained through analysis, for example, text fragments with similar contents such as 'restaurant a in building B' or 'arrival at restaurant a 2-level building B' appear in news for many times, and the mutual attachment relationship between the two POIs can be determined after semantic analysis processing.
In an optional embodiment, the POI presence state acquiring device may establish a POI presence information feature library according to the acquired network information feature data associated with the POI, so as to read the network information feature data associated with the target POI from the POI presence information feature library when needed.
And S102, converting the acquired network information feature data into corresponding present feature vectors.
The process of converting the network information feature data into the corresponding present feature vector can be regarded as a process of normalization processing. According to the obtained characteristic data of each kind of network information or the combination of at least two kinds of network information characteristic data, the present characteristic vector values of different dimensions of the target POI can be determined, the set present characteristic vector values of all dimensions form the present characteristic vector of the target POI, if the POI present state obtaining equipment obtains a certain kind of network information characteristic data, the present characteristic vector value of the corresponding dimension can be determined according to the obtained network information characteristic data, in the concrete implementation, the characteristic vector value corresponding to each kind of network information characteristic data can be determined through the preset comparison table of the network information characteristic data and the characteristic vector value of the corresponding dimension, the following table 1 is taken as an example, the characteristic vector values corresponding to 8 kinds of network information characteristic data can be obtained, the (-1, 1, -1, 0, 0, 1, 1, 1, 0) can be the present characteristic vector of the actually obtained target POI, in other alternative embodiments, part of the comparison relationships may be used to determine the present feature vector of the target POI, or more comparison relationships may be added to the comparison table to determine the present feature vector of the target POI. If the POI presence state acquiring device does not acquire a certain network information feature data of the target POI, the presence feature vector value of the corresponding dimension of the network information feature data may be set to a default value, for example, 0.
Figure BDA0001493385390000061
Figure BDA0001493385390000071
Table 1: comparison table of network information characteristic data and characteristic vector value
In an optional embodiment, if the acquired network information feature data associated with the target POI further includes network information feature data associated with the target POI and network information feature data associated with other POIs having an attachment relationship with the target POI, then in the same manner, the present feature vectors of the other POIs having an attachment relationship with the target POI may be obtained according to the network information feature data associated with the other POIs having an attachment relationship with the target POI.
And S103, inputting the converted situation characteristic vector of the target POI into a situation state judgment model to obtain the situation state of the target POI.
The POI situation state acquisition equipment inputs the converted situation feature vector of the target POI into the situation state judgment model, and then the situation state of the target POI can be judged and obtained.
It should be noted that, in the embodiment of the present application, the situation state determination model is obtained by training according to a plurality of POI with situation states labeled and acquired situation feature vectors corresponding to the POIs with situation states labeled, and the situation states corresponding to the POIs with situation states labeled are training samples. The situation feature vector corresponding to each POI labeled with a situation state may be obtained by transforming, in the same manner as in S101 to S102 in this embodiment, the network information feature data of each POI labeled with a situation state. It can be understood that, through the continuously trained present state judgment model, the present feature vectors of the plurality of sample POIs are substituted into the present state judgment model to perform the present state judgment result, so that the present state judgment result can continuously approach the labeling results of the plurality of sample POIs, and finally, the present state judgment model with the judgment result most approaching the labeling result of the sample POIs is determined as the trained present state judgment model.
In specific implementation, the present state judgment model may adopt machine learning models such as GBDT (Gradient Boosting Decision Tree), XGboost machine learning function library (XGboost) or DNN (Deep Neural network).
In an alternative embodiment, the present state judgment model may be obtained by fusing at least two present state judgment submodels, where each present state judgment submodel corresponds to at least one present feature vector. The at least two display state judgment submodels may also be combinedAnd a fusion function to form a present state judgment model, wherein the fusion function may include a plurality of fusion parameters, and each present state judgment sub-model may correspond to one fusion parameter in the present state judgment model. The present state judgment submodel may also adopt any one or more of GBDT, XGboost, DNN and other machine learning models. The relationship between the present status judging model and at least two present status judging submodels can be as shown in FIG. 3, wherein the present status judging model F (x) is composed of a plurality of present status judging submodels F (x)n(xn) And a fusion parameter a corresponding to each of the present state judgment submodelsnThe formed fusion functions are jointly formed, and the forming mode can be linear or nonlinear, and the following are exemplified:
F(x)＝a1·F1(x1)+a2·F2(x2)+a3·F3(x3).....+an·Fn(xn)
wherein F (x) represents a present state judgment model, Fn(xn) A situation feature vector representing a certain dimension and a corresponding situation state judgment submodel, anAnd a fusion parameter corresponding to the present state judgment submodel is indicated.
The POI present state acquiring device may train the corresponding present state judgment submodel according to the plurality of present state labeled POIs (i.e., sample POIs) and the acquired present feature vectors of the respective present state labeled POIs, and it is understood that the present state judgment submodel that is continuously trained substitutes the present feature vectors of the plurality of sample POIs into the present state judgment submodel to perform the judgment result of the present state to be able to continuously approximate the marking result of the plurality of sample POIs, and finally determines the present state judgment submodel whose judgment result most approximates the sample POI marking result as the trained present state judgment submodel.
And then the POI situation state acquisition equipment trains the fusion parameters corresponding to each situation state judgment sub-model in the situation state judgment model according to the POIs marked with the situation states, the acquired situation characteristic vectors corresponding to the POIs marked with the situation states and the trained at least two situation state judgment sub-models. And the trained situation state judgment model is formed according to the trained at least two situation state judgment submodels and the trained fusion parameters corresponding to each situation state judgment submodel.
And S104, outputting the current state of the target POI.
Specifically, the POI present state acquiring device may mark the present state of the target POI in the electronic map data, so that when the target POI is displayed on the electronic map, the present state of the target POI may be displayed at the same time to prompt the user. For example, as shown in fig. 4, when the user opens the electronic map, after a specific POI zhou ji duck blood vermicelli soup (lane-fixing shop) is clicked on the map or selected by the POI search box, the electronic map displays the relevant information of the POI, at this time, the present state of the POI may be prompted to the user (the "merchant close" information item in the figure), and in addition, when the user moves the cursor above the target POI, the relevant information including the present state of the POI may be displayed to the user.
In another embodiment, when the POI presence state acquiring device queries a POI according to a POI retrieval request, if the POI meeting the POI retrieval request includes the target POI, the returned POI query result carries other POIs excluding the target POI from the POIs, or sets the display ranking of the target POI in the POIs in the returned POI query result at the end. For example, as shown in fig. 5, a user inputs a search condition of "zhou ji duck blood vermicelli soup" in a search box, submits a POI search request to a POI present state acquisition device, the POI present state acquisition device performs a POI query in electronic map data according to the search request, and a target POI "zhou ji duck blood vermicelli soup (lane-fixing shop) also satisfies the search condition, but the target POI is also queried because the present condition is" closed ", so that the target POI may be excluded from the query result provided below the search box as shown in fig. 5, only other POIs satisfying the search condition may be provided, in other embodiments, the display ranking of the target POI may be set at the end of all POIs satisfying the search condition, and the present state of the target POI may be directly displayed in the query result. By changing the feedback result of the POI inquiry in the mode, the situation that the user finds the POI in the closed or unavailable state on the map and inconvenience is caused can be effectively avoided.
In other optional embodiments, the POI present status acquiring device may further enable the latter to mark the present status of the target POI in the electronic map data or change the feedback result of the POI query by outputting the present status of the target POI to the electronic map database or the electronic map data provider.
The POI present state acquiring device in this embodiment can rapidly determine the present state of the target POI according to the network information feature data and the present state determination model associated with the target POI, without requiring inefficient manual data acquisition, reporting, and auditing processes, thereby effectively ensuring the POI present state in the electronic map data.
Referring to fig. 6 and 7, fig. 7 is a schematic diagram of a logic framework implemented by a method for acquiring a POI presence state according to another embodiment of the present invention, and fig. 6 is a schematic flow chart of the method for acquiring a POI presence state implemented under the logic framework. As shown in fig. 6, the method of the embodiment of the present invention may include the following steps S201 to S208.
S201, acquiring the current states of the POI labels of the multiple samples.
The plurality of sample POIs, i.e., POIs marked with the current situation, may be manually marked or provided by a third party.
S202, network information characteristic data related to a plurality of sample POIs are obtained.
As shown in fig. 7, in this embodiment, the network information feature data associated with the sample POI includes WIFI access data corresponding to the sample POI, user click behavior data for the sample POI, and user comment data. In other optional embodiments, the network information characteristic data associated with the sample POI may include more contents, and reference may be specifically made to the description about the network information characteristic data in S101 in the foregoing embodiment, which is not described in detail in this embodiment.
S203, converting the network information characteristic data of the sample POI into a corresponding present characteristic vector.
According to the acquired feature data of each type of network information or the combination of at least two types of network information feature data, the situation feature vector values of different dimensions of the sample POI can be determined, the situation feature vector values of all the set dimensions form the situation feature vector of the sample POI, if the POI situation state acquisition equipment acquires certain type of network information feature data, the situation feature vector value of the corresponding dimension can be determined according to the acquired network information feature data, and in specific implementation, the feature vector value corresponding to each type of network information feature data can be determined through a preset comparison table of the network information feature data and the feature vector value of the corresponding dimension.
And S204, obtaining the situation characteristic vectors of the sample POIs according to the situation states marked by the sample POIs and the conversion, and training a situation state judgment model.
In specific implementation, the present state judgment model may adopt machine learning models such as GBDT, XGboost or DNN. Training the situation state judgment model can be regarded as a process of continuously optimizing model parameters of the situation state judgment model, and it can be understood that through the continuously trained situation state judgment model, the situation characteristic vectors of a plurality of sample POIs obtained through conversion are substituted into the situation state judgment model to carry out the situation state judgment result, which can continuously approximate the situation state result labeled by the plurality of sample POIs, and finally the situation state judgment model of which the judgment result is most approximate to the labeled result of the training sample is determined as the trained situation state judgment model.
In an optional embodiment, if the situation state determination model is composed of a plurality of situation state determination submodels and a fusion function composed of fusion parameters corresponding to each situation state determination submodel as shown in fig. 3, the POI situation state acquisition device may train the corresponding situation state determination submodel according to the situation states labeled by the plurality of sample POIs and the acquired situation feature vectors of the sample POIs, it may be understood that the situation state determination submodels continuously trained are substituted with the situation state determination submodels by the continuously trained situation state determination submodels, the determination result of the situation state can be continuously approximated to the labeling result of the plurality of sample POIs, and finally the situation state determination submodel of which the determination result is most approximated to the POI labeling result is determined as the trained situation state determination submodel. And then the POI situation state acquisition equipment trains the fusion parameters corresponding to each situation state judgment sub-model in the situation state judgment model according to the POIs marked with the situation states, the acquired situation characteristic vectors corresponding to the POIs marked with the situation states and the trained at least two situation state judgment sub-models. And therefore, the trained situation state judgment model is formed according to the trained at least two situation state judgment submodels and the trained fusion parameters corresponding to each situation state judgment submodel.
The above S201 to S204 are training processes of the present state judgment model, and the training processes may be completed independently from the following descriptions of S205 to S208, or may be sequentially performed according to the sequence of the present embodiment. The training process can be completed off line, and also can be performed on-line training according to the marking result of the sample POI and the network information characteristic data of the sample POI which are determined or updated in real time, and continuously retraining and optimizing the situation state judgment model in real time.
S205, acquiring the associated network information characteristic data of the target POI.
The target POI in this embodiment is all POIs in the specific POI set, that is, the current states of all POIs in the specific POI set can be updated once by executing S205-S207 in this embodiment according to the preset period, so that POIs whose current states change are found in time.
As shown in fig. 6, the network information feature data associated with the target POI in the present embodiment includes WIFI access data corresponding to the target POI, user click behavior data for the target POI, and user comment data. In other optional embodiments, the network information feature data associated with the sample POI may include more contents, and reference may be specifically made to the description about the network information feature data in S101 in the foregoing embodiment, which is not described in detail in this embodiment.
And S206, converting the network information characteristic data of the target POI into a corresponding present characteristic vector.
According to the obtained each kind of network information characteristic data or the combination of at least two kinds of network information characteristic data, the present feature vector values of different dimensions of the target POI can be determined, the set present feature vector values of all dimensions form the present feature vector of the target POI, if the POI present state obtaining equipment obtains a certain kind of network information characteristic data, the present feature vector value of the corresponding dimension can be determined according to the obtained network information characteristic data, and in the specific implementation, the feature vector value corresponding to each kind of network information characteristic data can be determined through a preset network information characteristic data and a feature vector value comparison table of the corresponding dimension.
And S207, inputting the situation characteristic vector of the target POI into the trained situation state judgment model to obtain the situation state of the target POI.
And S208, outputting the current state of the target POI to an application scene.
The application scenario may be, for example, an electronic map application scenario shown in fig. 4 or fig. 5, and may also, for example, disclose or broadcast the POI whose presence status changes, or change a user route plan or a navigation route according to the POI whose presence status changes.
In the embodiment, the situation state judgment model is trained by using the situation states marked by the multiple sample POIs and the situation characteristic vectors of the multiple sample POIs, so that the situation state of the target POI can be judged by the trained situation state judgment model, inefficient manual data acquisition, reporting and auditing processes are not needed, and the POI situation in the electronic map data is effectively ensured.
Fig. 8 is a schematic flow chart illustrating a method for acquiring a POI present status according to another embodiment of the present invention. As shown in fig. 8, the method of the embodiment of the present invention may include the following steps S301 to S309.
S301, collecting the network media text information in the first time period.
The first time period in this embodiment may be a window of time of the last 2 weeks, 3 weeks, or the last month. The POI present state acquisition equipment can acquire a large amount of network media text information through various media channels such as information webpage data in webpage media, information data of news APP, articles, trends and messages of social network media and the like. In an alternative embodiment, the POI present status acquiring device may re-collect the network media text information according to a set period.
S302, extracting a media text fragment related to the POI name from the collected network media text information according to the full POI name set and the first entity identification model.
The full-amount POI name set in this embodiment may be all POI names in the electronic map data, that is, the POI present status acquiring device screens out a media text segment associated with a POI from the collected network media text information in the first time period according to all POI names in the electronic map data.
The first entity recognition model comprises an AC automaton, a Conditional Random Field (CRF) or a bidirectional Long-Short Term Memory conditional random field neural network (BI-LSTM-CRF), namely, any one of the following three ways is adopted to screen out media text fragments related to POI from the collected network media text information in the first time period:
1) and matching all POI names in the electronic map data with words appearing in the collected network media text information through an AC automatic machine model, so that a media text segment containing the POI names can be obtained. Similarly, a Darts (Double-ARray Trie System) model can also be used.
2) And training the conditional random field practice recognition model in advance according to the full POI name set and the media text fragments marked with the related POIs, and judging whether the collected network media text information is the media text fragment associated with the POI or not through the trained conditional random field practice recognition model, compared with the mode 1), the situation of semantic ambiguity can be effectively avoided, and the recognition precision is improved.
3) Corresponding word feature vectors are respectively generated by words and phrases appearing in the media text segments of the multiple marked related POIs, and a BI-LSTM-CRF model is trained in advance according to the word feature vectors appearing in the full POI name set and the media text segments of the multiple marked related POIs, so that the recognition accuracy of an entity recognition model on the text prefix and suffix contents related to the actual POI names in text information can be effectively improved, and the recognition accuracy of the media text segments related to the POIs is integrally improved.
And further analyzing the data to obtain media segments related to the current state of the POI, and judging the current state of the POI, for example, if a news content is fit up for a dead passenger at one floor of a certain building, the target POI related to the news data is the dead passenger at the floor of the building, and the news data can be obviously used as network media data related to the target POI
S303, acquiring a media text fragment related to the target POI from the media text fragments related to the POI name according to the target POI name.
Step S302 may screen out all media text segments associated with the POI from the collected network media text information in the first time period, and this step further obtains the media text segments related to the target POI from the network media text information according to the name of the target POI.
S304, in the media text fragments related to the target POI, natural language processing is carried out on the media text fragments simultaneously containing the target POI name and other POI names, and the attachment relation between the target POI and other POI is obtained.
That is, it is determined that two POIs having an attachment relationship need to appear in the same media text segment first, and then the POI presence state acquiring device performs natural language processing on the media text segment in which the two POIs have the same name, for example, text segments with similar contents such as "restaurant a in building B" or "go to restaurant a 2-level in building B" appear in news many times, and after semantic analysis processing, it can be determined that the attachment relationship exists between the two POIs. Keywords used for keyword matching may be, for example, "… … in … …", "in … …", "… … in … …", and so on.
S305, extracting media text fragments related to other POIs which have an attachment relation with the target POI from the collected network media text information according to the names of the other POIs which have an attachment relation with the target POI.
S306, respectively carrying out natural language processing on the media text fragments related to the target POI and the media text fragments related to other POIs having an attachment relation with the target POI to obtain the current data of the target POI and the current data of the other POIs having the attachment relation with the target POI as network media data related to the target POI.
That is, natural language processing is performed on the media text segment related to the target POI, such as semantic analysis and keyword matching, so that the present data of the target POI can be obtained through analysis. The keywords that reflect the present situation generally include: and performing semantic analysis on sentences comprising the names of the target POI and the keywords reflecting the current situation state to obtain the current situation data of the target POI. Similarly, the media text segments related to other POIs with an attachment relation with the target POI are subjected to natural language processing, so that the situation data of the other POIs with the attachment relation with the target POI can be obtained through analysis.
And S307, converting the network information feature data of the target POI into a corresponding situation feature vector.
In this embodiment, the network media data to be associated with the target POI includes the situation data of the target POI and the situation data of other POIs having an attachment relationship with the target POI, and is converted into corresponding situation feature vectors. The process of converting the network media data into the corresponding present feature vector can be regarded as a process of normalization processing. According to the acquired situation data of each different situation or the combination of at least two kinds of situation data, situation feature vector values of different dimensions of the target POI can be determined, the situation feature vector values of all the dimensions set form a situation feature vector of the target POI, and in the concrete implementation, the feature vector value corresponding to each kind of network information feature data can be determined through a preset situation data and corresponding dimension feature vector value comparison table.
And S308, inputting the situation characteristic vector of the target POI into the trained situation state judgment model to obtain the situation state of the target POI.
And S309, outputting the current state of the target POI to an application scene.
The application scenario may be, for example, an electronic map application scenario shown in fig. 4 or fig. 5, and may also, for example, disclose or broadcast the POI whose presence status changes, or change a user route plan or a navigation route according to the POI whose presence status changes.
The embodiment analyzes the network media data to obtain the network media data which are associated with the target POI and comprise the current data of the target POI and the current data of other POIs which have an attachment relation with the target POI, and judges the current state of the target POI by combining the trained current state judgment model, so that the inefficient manual data acquisition, reporting and auditing processes are not needed any more, and the POI current state in the electronic map data is effectively ensured.
Referring to fig. 9, a schematic structural diagram of a POI presence status acquiring apparatus is provided in an embodiment of the present invention. As shown in fig. 9, the POI presence status acquiring apparatus according to the embodiment of the present invention may include:
a network information obtaining module 910, configured to obtain network information feature data associated with the target POI.
The network information characteristic data associated with the target POI comprises access data of a wireless access point corresponding to the target POI, network user behavior data associated with the target POI or network media data associated with the target POI.
The network information obtaining module 910 obtains network media data associated with a target POI, including:
collecting network media text information in a first time period;
extracting a media text fragment related to a target POI from the collected network media text information according to the target POI name, specifically, extracting the media text fragment related to the POI name from the collected network media text information according to a full POI name set and a first entity recognition model, wherein the first entity recognition model comprises an AC automaton, a conditional random field entity recognition model or a two-way long-short term memory conditional random field neural network model, and then acquiring the media text fragment related to the target POI from the media text fragment related to the POI name according to the target POI name, wherein the first entity recognition model can be obtained by training according to the full POI name set and a plurality of media text fragments marked with related POI as training samples;
and carrying out natural language processing on the media text segments related to the target POI to obtain the current data of the target POI as network media data associated with the target POI. Specifically, the method may include performing natural language processing on a media text segment including a target POI name and other POI names in a media text segment related to the target POI to obtain an attachment relationship between the target POI and other POIs, extracting media text segments related to other POIs having an attachment relationship with the target POI from the collected network media text information according to names of other POIs having an attachment relationship with the target POI, and further performing natural language processing on the media text segments related to the target POI and the media text segments related to other POIs having an attachment relationship with the target POI to obtain situation data of the target POI and situation data of the other POIs having an attachment relationship with the target POI as network media data related to the target POI.
In an optional embodiment, the network information obtaining module 910 may further use, according to an attachment relationship between the target POI and other POIs, network information feature data of other POIs having an attachment relationship with the target POI as network information feature data associated with the target POI, that is, the network information feature data associated with the target POI may further include the network information feature data related to the target POI and the network information feature data related to other POIs having an attachment relationship with the target POI.
The present feature converting module 920 is configured to convert the acquired network information feature data into corresponding present feature vectors.
The present state determination module 930 is configured to input the converted present feature vector of the target POI into the present state determination model to obtain a present state of the target POI.
The present state determination module 930 may determine the present state of the target POI by inputting the converted present feature vector of the target POI into the present state determination model.
In specific implementation, the present state judgment model may adopt GBDT (Gradient Boosting Decision Tree), XGboost machine learning function library, DNN (Deep Neural network), or other machine learning models.
In an alternative embodiment, the present state judgment model may be obtained by fusing at least two present state judgment submodels, wherein each present state judgment submodel corresponds to at least one present feature vector.
A status of present output module 940, configured to output the status of present of the target POI.
Specifically, the present state output module 940 may be configured to mark the present state of the target POI in the electronic map data; or when the POI search request is used for carrying out the POI search, if the POI meeting the POI search request comprises the target POI, carrying other POI excluding the target POI in the returned POI search result, or setting the display sequence of the target POI in the returned POI search result at the last.
In an optional embodiment, the POI present status acquiring device may further include:
the judgment model training module 950 is configured to train the situation state judgment model according to the multiple POIs marked with situation states and the acquired situation feature vectors corresponding to the POIs marked with situation states. That is, the situation state determination model may be obtained by the determination model training module 950 through training according to the multiple POI already labeled with the situation state and the obtained situation feature vector corresponding to each POI already labeled with the situation state, where the situation state corresponding to the multiple POI already labeled with the situation state is the training sample. The present feature vector corresponding to each POI labeled with a present state may be obtained by the present feature conversion module 920 according to the network information feature data of each POI labeled with a present state. It can be understood that through the continuously trained situation state judgment model, the situation characteristic vectors of the multiple sample POIs are substituted into the situation state judgment model to perform the situation state judgment result, so that the situation state judgment result can be continuously approximated to the labeling results of the multiple sample POIs, and finally the situation state judgment model of which the judgment result is most approximated to the labeling results of the sample POIs is determined as the trained situation state judgment model.
If the situation state judgment model is obtained by fusing at least two situation state judgment submodels, the judgment model training module 950 may train the corresponding situation state judgment submodels according to the multiple POI (i.e. sample POI) marked with situation states and the obtained situation feature vectors of the POI marked with situation states, it can be understood that the situation state judgment submodels are continuously trained, the situation feature vectors of the multiple sample POIs are substituted into the situation state judgment submodels to perform the judgment result of the situation state to continuously approximate the marking result of the multiple sample POIs, and finally the situation state judgment submodel of which the judgment result most approximates the marking result of the sample POI is determined as the trained situation state judgment submodel;
then, the judgment model training module 950 trains the fusion parameters corresponding to each of the situation state judgment submodels in the situation state judgment model according to the plurality of situation state labeled POIs, the acquired situation feature vectors corresponding to the POIs labeled with the situation states, and the trained at least two situation state judgment submodels. And obtaining the trained situation state judgment model according to the trained at least two situation state judgment submodels and the trained fusion parameters corresponding to each situation state judgment submodel.
The POI present state acquiring device in this embodiment can rapidly determine the present state of the target POI according to the network information feature data and the present state determination model associated with the target POI, without requiring inefficient manual data acquisition, reporting, and auditing processes, thereby effectively ensuring the POI present state in the electronic map data.
An embodiment of the present invention further provides a computer storage medium, where multiple instructions may be stored in the computer storage medium, where the instructions are suitable for being loaded by a processor and being executed in the method steps in the embodiments shown in fig. 1 to fig. 8, and specific execution processes may refer to specific descriptions in the embodiments shown in fig. 1 to fig. 8, which are not described herein again.
Fig. 10 is a schematic structural diagram of a server according to an embodiment of the present invention. As shown in fig. 10, the server 1000 may include: at least one processor 1001, such as a CPU, at least one network interface 1004, a user interface 1003, memory 1005, at least one communication bus 1002. The communication bus 1002 is used to implement connection communication among these components. The network interface 1004 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface). The memory 1005 may be a high-speed RAM memory or a non-volatile memory (non-volatile memory), such as at least one disk memory. The memory 1005 may optionally be at least one memory device located remotely from the processor 1001. As shown in fig. 10, a memory 1005, which is a kind of computer storage medium, may include therein an operating system, a network communication module, and a POI presence state acquisition program.
In a specific implementation, the server in the embodiment of the present invention may be deployed on a physical cluster composed of hundreds of physical machines by using Hadoop (a software framework capable of performing distributed processing on a large amount of data) and Hive (a data warehouse tool based on Hadoop) large data processing technologies. The Hive database can be used for summarizing and storing collected various network information characteristic data in the cluster, then Mapreduce (a distributed computing framework model) is used for performing characteristic extraction and preliminary characteristic engineering computation, and after extraction, the collected characteristic data is written into an unstructured database (such as a BDB (distributed computing architecture) DB database) to be stored for use.
In the server 1000 shown in fig. 10, the network interface 1004 is mainly used for data communication with an external device, such as acquiring network information feature data associated with a target POI from the external device and outputting a present state of the POI to the external device; the processor 1001 may be configured to invoke the traffic information obtaining application stored in the memory 1005, and specifically execute the following operations:
acquiring network information characteristic data associated with a target POI;
converting the acquired network information characteristic data into corresponding situation characteristic vectors;
inputting the converted situation characteristic vector of the target POI into a situation state judgment model to obtain the situation state of the target POI;
and outputting the current state of the target POI.
In one embodiment, the network information characteristic data associated with the target POI comprises network information characteristic data related to the target POI and network information characteristic data related to other POIs having an attachment relation with the target POI.
In one embodiment, the network information characteristic data associated with the target POI comprises access data of a wireless access point corresponding to the target POI, network user behavior data associated with the target POI, or network media data associated with the target POI.
In an embodiment, the executing of the processor 1001 to acquire the network media data associated with the target POI specifically includes:
collecting network media text information in a first time period;
extracting a media text fragment related to the target POI from the collected network media text information according to the name of the target POI;
and carrying out natural language processing on the media text segments related to the target POI to obtain the current data of the target POI as network media data associated with the target POI.
In one embodiment, the processor 1001 extracts a media text segment related to the target POI from the collected network media text information according to the name of the target POI specifically includes:
extracting media text fragments related to the POI names from the collected network media text information according to a full POI name set and a first entity recognition model, wherein the first entity recognition model comprises an AC automaton, a conditional random field entity recognition model or a bidirectional long-short term memory conditional random field neural network model;
and acquiring a media text fragment related to the target POI from the media text fragments related to the POI name according to the target POI name.
In one embodiment, the processor 1001, before executing the step of extracting the media text segment related to the POI name from the collected network media text information according to the full POI name set and the first entity recognition model, further executes the following operations:
and training the first entity recognition model according to the full POI name set and the media text segments of the plurality of marked related POIs.
In one embodiment, the performing, by the processor 1001, natural language processing on the media text segment related to the target POI to obtain the presence data of the target POI as network media data associated with the target POI specifically includes:
in a media text fragment related to a target POI, performing natural language processing on the media text fragment simultaneously containing the target POI name and other POI names to obtain an attachment relation between the target POI and other POIs;
extracting media text fragments related to other POIs which have an attachment relation with the target POI from the collected network media text information according to the names of the other POIs which have the attachment relation with the target POI;
and respectively carrying out natural language processing on the media text fragments related to the target POI and the media text fragments related to other POIs having an attachment relation with the target POI to obtain the current data of the target POI and the current data of the other POIs having the attachment relation with the target POI as network media data related to the target POI.
In one embodiment, before the processor 1001 inputs the feature vector of the target POI obtained by conversion into the situational state judgment model to obtain the situational state of the target POI, the following operations are further performed:
and training the situation state judgment model according to the POIs marked with the situation states and the acquired situation characteristic vectors corresponding to the POIs marked with the situation states.
In one embodiment, the present state judgment model is obtained by fusing at least two present state judgment submodels, wherein each present state judgment submodel corresponds to at least one present feature vector.
Further, before the processor 1001 inputs the converted present feature vector of the target POI into the present state judgment model to obtain the present state of the target POI, the processor further performs the following operations:
training a corresponding situation state judgment sub-model according to the POIs marked with the situation states and the acquired situation characteristic vectors of the POIs marked with the situation states;
and training the fusion parameters corresponding to each situation state judgment sub-model in the situation state judgment model according to the POIs marked with the situation states, the acquired situation characteristic vectors corresponding to the POIs marked with the situation states and the trained at least two situation state judgment sub-models.
In one embodiment, the executing of the processor 1001 to output the present state of the target POI specifically includes:
and marking the current state of the target POI in the electronic map data.
In another embodiment, the executing of the processor 1001 to output the present state of the target POI specifically includes:
when POI query is carried out according to a POI retrieval request, if a plurality of POIs meeting the POI retrieval request comprise the target POI, carrying other POIs excluding the target POI in the POI query results returned, or setting the display sequence of the target POI in the POI query results returned at the end.
The server in this embodiment can rapidly determine the current state of the target POI according to the network information feature data and the current state determination model associated with the target POI, and inefficient manual data acquisition, reporting, and auditing processes are no longer required, so that the POI current state in the electronic map data is effectively ensured.
It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.
The above disclosure is only for the purpose of illustrating the preferred embodiments of the present invention, and it is therefore to be understood that the invention is not limited by the scope of the appended claims.

Claims (13)

1. A POI present state acquisition method is characterized by comprising the following steps:
collecting network media text information in a first time period;
extracting a media text fragment related to the target POI from the collected network media text information according to the name of the target POI;
in a media text fragment related to a target POI, performing natural language processing on the media text fragment simultaneously containing the target POI name and other POI names to obtain an attachment relation between the target POI and other POIs;
extracting media text fragments related to other POIs which have an attachment relation with the target POI from the collected network media text information according to the names of the other POIs which have an attachment relation with the target POI;
respectively carrying out natural language processing on the media text segments related to the target POI and the media text segments related to other POIs which have an attachment relation with the target POI to obtain the current data of the target POI and the current data of the other POIs which have the attachment relation with the target POI as network media data associated with the target POI;
converting network information characteristic data associated with a target POI into a corresponding situation characteristic vector, wherein the network information characteristic data associated with the target POI comprises network media data associated with the target POI;
inputting the converted situation characteristic vector of the target POI into a situation state judgment model to obtain the situation state of the target POI;
and outputting the current state of the target POI.
2. The method of claim 1, wherein the network information characteristic data associated with the target POI comprises network information characteristic data related to the target POI and network information characteristic data related to other POIs having an attachment relationship with the target POI.
3. The method of claim 1, wherein the network information characteristic data associated with the target POI further comprises access data of a wireless access point corresponding to the target POI or network user behavior data associated with the target POI.
4. The method of claim 1, wherein said extracting a media text segment related to a target POI from said collected network media text information according to the target POI name comprises:
extracting media text fragments related to the POI names from the collected network media text information according to a full POI name set and a first entity recognition model, wherein the first entity recognition model comprises an AC automaton, a conditional random field entity recognition model or a two-way long-short term memory conditional random field neural network model;
and acquiring a media text fragment related to the target POI from the media text fragments related to the POI name according to the target POI name.
5. The method of claim 4, wherein said extracting media text segments related to POI names from said collected network media text information according to a full set of POI names and a first entity recognition model further comprises:
and training the first entity recognition model according to the full POI name set and the media text segments of the plurality of marked related POIs.
6. The method of claim 1, wherein the step of inputting the converted present feature vector of the target POI into the present state judgment model further comprises, before obtaining the present state of the target POI:
and training the situation state judgment model according to the POIs marked with the situation states and the acquired situation characteristic vectors corresponding to the POIs marked with the situation states.
7. The method of claim 1, wherein the presence state judgment model is fused from at least two presence state judgment submodels, wherein each presence state judgment submodel corresponds to at least one presence feature vector.
8. The method as claimed in claim 7, wherein the step of inputting the converted present feature vector of the target POI into the present state judgment model further comprises:
training a corresponding situation state judgment sub-model according to the POIs marked with the situation states and the acquired situation characteristic vectors of the POIs marked with the situation states;
and training the fusion parameters corresponding to each situation state judgment sub-model in the situation state judgment model according to the POIs marked with the situation states, the acquired situation characteristic vectors corresponding to the POIs marked with the situation states and the trained at least two situation state judgment sub-models.
9. The method of any one of claims 1-8, wherein the outputting the present state of the target POI comprises:
and marking the current situation of the target POI in the electronic map data.
10. The method of any one of claims 1-8, wherein the outputting the present state of the target POI comprises:
when POI query is carried out according to a POI retrieval request, if a plurality of POIs meeting the POI retrieval request comprise the target POI, carrying other POIs excluding the target POI in the POI query results returned, or setting the display sequence of the target POI in the POI query results returned at the end.
11. A POI presence state acquiring apparatus, characterized in that the apparatus comprises:
the network information acquisition module is used for collecting network media text information in a first time period; extracting a media text fragment related to the target POI from the collected network media text information according to the name of the target POI; in a media text fragment related to a target POI, performing natural language processing on the media text fragment simultaneously containing the target POI name and other POI names to obtain an attachment relation between the target POI and other POIs; extracting media text fragments related to other POIs which have an attachment relation with the target POI from the collected network media text information according to the names of the other POIs which have an attachment relation with the target POI; respectively carrying out natural language processing on the media text fragments related to the target POI and the media text fragments related to other POIs which have an attachment relation with the target POI to obtain the current data of the target POI and the current data of the other POIs which have the attachment relation with the target POI as network media data related to the target POI;
the situation feature conversion module is used for converting network information feature data associated with a target POI into a corresponding situation feature vector, wherein the network information feature data associated with the target POI comprise network media data associated with the target POI;
the situation state judgment module is used for inputting the situation characteristic vector of the target POI obtained by conversion into the situation state judgment model to obtain the situation state of the target POI;
and the current state output module is used for outputting the current state of the target POI.
12. A computer storage medium having stored thereon a plurality of instructions adapted to be loaded by a processor and to perform the steps of:
collecting network media text information in a first time period;
extracting media text segments related to the target POI from the collected network media text information according to the name of the target POI;
in a media text fragment related to a target POI, performing natural language processing on the media text fragment simultaneously containing the target POI name and other POI names to obtain an attachment relation between the target POI and other POIs;
extracting media text fragments related to other POIs which have an attachment relation with the target POI from the collected network media text information according to the names of the other POIs which have an attachment relation with the target POI;
respectively carrying out natural language processing on the media text fragments related to the target POI and the media text fragments related to other POIs which have an attachment relation with the target POI to obtain the current data of the target POI and the current data of the other POIs which have the attachment relation with the target POI as network media data related to the target POI;
converting network information characteristic data associated with a target POI into a corresponding situation characteristic vector, wherein the network information characteristic data associated with the target POI comprises network media data associated with the target POI;
inputting the converted situation characteristic vector of the target POI into a situation state judgment model to obtain the situation state of the target POI;
and outputting the current state of the target POI.
13. A server, comprising: a processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and to perform the steps of:
collecting network media text information in a first time period;
extracting a media text fragment related to the target POI from the collected network media text information according to the name of the target POI;
in a media text fragment related to a target POI, performing natural language processing on the media text fragment simultaneously containing the target POI name and other POI names to obtain an attachment relation between the target POI and other POIs;
extracting media text fragments related to other POIs which have an attachment relation with the target POI from the collected network media text information according to the names of the other POIs which have an attachment relation with the target POI;
respectively carrying out natural language processing on the media text fragments related to the target POI and the media text fragments related to other POIs which have an attachment relation with the target POI to obtain the current data of the target POI and the current data of the other POIs which have the attachment relation with the target POI as network media data related to the target POI;
converting network information characteristic data associated with a target POI into a corresponding situation characteristic vector, wherein the network information characteristic data associated with the target POI comprises network media data associated with the target POI;
inputting the converted situation characteristic vector of the target POI into a situation state judgment model to obtain the situation state of the target POI;
and outputting the current state of the target POI.
CN201711260201.5A 2017-12-04 2017-12-04 POI (Point of interest) situation acquisition method and equipment, storage medium and server thereof Active CN110019616B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711260201.5A CN110019616B (en) 2017-12-04 2017-12-04 POI (Point of interest) situation acquisition method and equipment, storage medium and server thereof

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711260201.5A CN110019616B (en) 2017-12-04 2017-12-04 POI (Point of interest) situation acquisition method and equipment, storage medium and server thereof

Publications (2)

Publication Number Publication Date
CN110019616A CN110019616A (en) 2019-07-16
CN110019616B true CN110019616B (en) 2022-07-12

Family

ID=67186530

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711260201.5A Active CN110019616B (en) 2017-12-04 2017-12-04 POI (Point of interest) situation acquisition method and equipment, storage medium and server thereof

Country Status (1)

Country Link
CN (1) CN110019616B (en)

Families Citing this family (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110347775B (en) * 2019-07-17 2022-11-08 北京百度网讯科技有限公司 Interest point state correction method, device, equipment and computer readable storage medium
CN111858787A (en) * 2019-09-24 2020-10-30 北京嘀嘀无限科技发展有限公司 POI information acquisition method and device
CN110851738B (en) * 2019-10-28 2021-03-19 百度在线网络技术（北京）有限公司 Method, device and equipment for acquiring POI state information and computer storage medium
CN110990728B (en) * 2019-12-03 2023-09-12 汉海信息技术(上海)有限公司 Method, device, equipment and storage medium for managing interest point information
CN111198916B (en) * 2020-01-03 2023-12-08 北京明略软件系统有限公司 Data transmission method and device, electronic equipment and storage medium
CN111782973A (en) * 2020-06-04 2020-10-16 汉海信息技术(上海)有限公司 Interest point state prediction method and device, electronic equipment and storage medium
CN111832483B (en) * 2020-07-14 2024-03-08 北京百度网讯科技有限公司 Point-of-interest validity identification method, device, equipment and storage medium
CN111860503A (en) * 2020-07-16 2020-10-30 北京奇虎科技有限公司 Information point validity identification method, device, equipment and storage medium
CN112330379B (en) * 2020-11-25 2023-10-31 税友软件集团股份有限公司 Invoice content generation method, invoice content generation system, electronic equipment and storage medium
CN113222492B (en) * 2021-03-29 2024-05-03 北京中交兴路信息科技有限公司 Method and device for discriminating type of vehicle driving line, storage medium and terminal
CN114860836B (en) * 2022-05-24 2023-03-10 北京百度网讯科技有限公司 Method, device, equipment and medium for mining failure interest points

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102788586A (en) * 2011-05-20 2012-11-21 广东瑞图万方科技股份有限公司 Method and system for incremental updating of navigation electronic map
CN103064953A (en) * 2012-12-28 2013-04-24 武汉大学 Point of interest (POI) updating method based on location check-in data
CN104866542A (en) * 2015-05-05 2015-08-26 腾讯科技（深圳）有限公司 POI data verification method and device
CN107133900A (en) * 2017-04-05 2017-09-05 中国农业大学 Urban land mixing utilizes feature grid computational methods and device

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9712964B2 (en) * 2015-12-21 2017-07-18 Amazon Technologies, Inc. Maintaining privacy in location-based operations

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102788586A (en) * 2011-05-20 2012-11-21 广东瑞图万方科技股份有限公司 Method and system for incremental updating of navigation electronic map
CN103064953A (en) * 2012-12-28 2013-04-24 武汉大学 Point of interest (POI) updating method based on location check-in data
CN104866542A (en) * 2015-05-05 2015-08-26 腾讯科技（深圳）有限公司 POI data verification method and device
CN107133900A (en) * 2017-04-05 2017-09-05 中国农业大学 Urban land mixing utilizes feature grid computational methods and device

Also Published As

Publication number Publication date
CN110019616A (en) 2019-07-16

Similar Documents

Publication Publication Date Title
CN110019616B (en) POI (Point of interest) situation acquisition method and equipment, storage medium and server thereof
CN110008300B (en) Method and device for determining alias of POI (Point of interest), computer equipment and storage medium
JP6894534B2 (en) Information processing method and terminal, computer storage medium
CN110825957A (en) Deep learning-based information recommendation method, device, equipment and storage medium
CN110825956A (en) Information flow recommendation method and device, computer equipment and storage medium
US20220245556A1 (en) Data distillery for signal detection
CN107885873B (en) Method and apparatus for outputting information
CN111159341B (en) Information recommendation method and device based on user investment and financial management preference
CN116127020A (en) Method for training generated large language model and searching method based on model
CN109739965B (en) Method, device and equipment for migrating cross-domain conversation strategy and readable storage medium
CN111325022B (en) Method and device for identifying hierarchical address
JP2007219655A (en) Facility information management system, facility information management method and facility information management program
CN110968801A (en) Real estate product searching method, storage medium and electronic device
EP3961426A2 (en) Method and apparatus for recommending document, electronic device and medium
CN111191133A (en) Service search processing method, device and equipment
CN112487109A (en) Entity relationship extraction method, terminal and computer readable storage medium
CN113742496B (en) Electric power knowledge learning system and method based on heterogeneous resource fusion
CN114417169A (en) Information recommendation optimization method, device, medium, and program product
CN117149140B (en) Method, device and related equipment for generating coded architecture information
CN116508004A (en) Method for point of interest information management, electronic device, and storage medium
CN117216393A (en) Information recommendation method, training method and device of information recommendation model and equipment
KR102532216B1 (en) Method for establishing ESG database with structured ESG data using ESG auxiliary tool and ESG service providing system performing the same
CN108520012B (en) Mobile internet user comment mining method based on machine learning
US20220327147A1 (en) Method for updating information of point of interest, electronic device and storage medium
CN115759100A (en) Data processing method, device, equipment and medium

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