CN113539270B

CN113539270B - Position identification method and device, electronic equipment and storage medium

Info

Publication number: CN113539270B
Application number: CN202110830026.9A
Authority: CN
Inventors: 姚雷; 杜新凯; 纪诚; 黄莹
Original assignee: Sunshine Insurance Group Co Ltd
Current assignee: Sunshine Insurance Group Co Ltd
Priority date: 2021-07-22
Filing date: 2021-07-22
Publication date: 2024-04-02
Anticipated expiration: 2041-07-22
Also published as: CN113539270A

Abstract

The application provides a position identification method, a position identification device, electronic equipment and a storage medium, wherein the method comprises the following steps: after voice data is acquired, converting the voice data into text information through a voice recognition technology; after inputting the text information into a pre-trained positioning model for carrying out fuzzy recognition on characters, recognizing the characters with the semantic similarity exceeding a preset threshold value with the place names in the text information through the positioning model, and taking the characters as target characters; for each target text, searching the geographic position information containing the target text from the corresponding relation between the text and the geographic position, so as to determine the geographic position in the geographic position information as the geographic position corresponding to the target text. According to the method, the accuracy of identifying the place names in the voice can be improved.

Description

Position identification method and device, electronic equipment and storage medium

Technical Field

The present application relates to the field of speech recognition technologies, and in particular, to a location recognition method, a location recognition device, an electronic device, and a storage medium.

Background

With the advent of speech recognition technology, the human-computer interaction mode of speech is gradually applied to more occasions, and when position recognition is performed, the position in the speech is also gradually recognized by adopting a speech input mode.

The inventors have found in research that in the prior art, speech is usually converted into text by means of speech recognition technology, and the place name in the text is determined by means of word-by-word or word-by-word comparison. In practical application, because the voice recognition technology is not mature enough, when noise, intonation change, uneven speaking modes and other interferences exist in voice data, errors are generated when the voice is converted into text by the voice recognition technology, so that partial words in the converted text are changed, and the accuracy is lower when the place names in the voice are recognized in a word-by-word or word-by-word comparison mode.

Disclosure of Invention

In view of this, the embodiments of the present application provide a position recognition method, apparatus, electronic device, and storage medium, so as to solve the problem that the accuracy of recognizing the place name in the voice is low.

In a first aspect, an embodiment of the present application provides a location identifying method, including:

after voice data is acquired, converting the voice data into text information through a voice recognition technology;

After inputting the text information into a pre-trained positioning model for carrying out fuzzy recognition on characters, recognizing the characters with the semantic similarity exceeding a preset threshold value with the place names in the text information through the positioning model, and taking the characters as target characters;

for each target text, searching the geographic position information containing the target text from the corresponding relation between the text and the geographic position, so as to determine the geographic position in the geographic position information as the geographic position corresponding to the target text.

In one possible embodiment, the positioning model is trained by:

after a plurality of sample voice data are acquired, converting the sample voice data into training texts by a voice recognition technology according to each sample voice data;

marking a first identifier for a first character used for representing the name of the place name in the training text, and marking a second identifier for a non-first character used for representing the name of the place name, so that the training text carrying the first identifier and the second identifier is used as a target training text;

a dataset comprising a plurality of target training texts is input into a bert+crf model to train the bert+crf model into the positioning model by means of supervised learning.

In one possible embodiment, inputting a dataset comprising a plurality of target training texts into a bert+crf model to train the bert+crf model into the positioning model by means of supervised learning, comprising:

respectively placing the target training files in the data set into a training set, a verification set and a test set according to a preset proportion;

after the training set, the verification set and the test set are respectively input into the BERT+CRF model, training the BERT+CRF model under the super parameters through at least one target training text in the training set and the first identifier and the second identifier carried by the at least one target training text aiming at each super parameter preset for the BERT+CRF model so as to obtain parameters of the BERT+CRF model under the super parameters; the BERT+CRF model identifies characters with semantic similarity exceeding a preset threshold value with the place name in the text information according to the parameters;

after obtaining the parameters, verifying the parameters of the BERT+CRF model under the super parameters through the verification set for each super parameter to obtain a first recognition rate of the model under the parameters; the first recognition rate is the success rate of recognizing characters with semantic similarity with place names exceeding a preset threshold value in the target training texts in the verification set by the BERT+CRF model under the parameter;

After the super-parameters of the BERT+CRF model are determined according to the first recognition rate, testing the parameters of the BERT+CRF model under the super-parameters through a testing set to obtain a second recognition rate of the model under the parameters; the second recognition rate is the success rate of recognizing characters with semantic similarity with place names exceeding a preset threshold value in the target training texts in the test set by the BERT+CRF model under the parameter;

judging whether the second recognition rate is larger than or equal to a preset recognition rate, and when the second recognition rate of the BERT+CRF model is larger than or equal to the preset recognition rate, taking the model as the positioning model to recognize characters with semantic similarity with place names exceeding a preset threshold value in the text information through the positioning model.

In one possible embodiment, the correspondence between the text and the geographic location is constructed by:

after at least one geographic position information containing a geographic position, a word used for representing the name of the geographic position and the corresponding relation between the word and the geographic position is obtained, for each geographic position in the geographic position information, the geographic position, at least one word or vocabulary used for representing the name of the geographic position and the corresponding relation between the at least one word or vocabulary and the geographic position are put into a geographic position set corresponding to the geographic position from at least one geographic position information corresponding to the geographic position;

And storing the geographical position set into the corresponding relation between the characters and the geographical positions.

In a possible embodiment, before obtaining at least one geographic location information including a geographic location, a text for representing a name of the geographic location, and a correspondence between the text and the geographic location, the method further includes:

for each marked word, taking the marked word as a preset place name, and setting a geographic position corresponding to the preset place name for each preset place name; the mark vocabulary is a vocabulary formed by characters carrying target marks, and the target marks comprise a first mark and a second mark;

and storing the preset place names, the geographic positions of the preset place names and the corresponding relation between the preset place names and the geographic positions into geographic position information corresponding to the preset place names aiming at each preset place name.

In a second aspect, embodiments of the present application further provide a location identifying apparatus, where the apparatus includes:

the conversion unit is used for converting the voice data into text information through a voice recognition technology after the voice data are acquired;

the positioning unit is used for identifying characters with semantic similarity exceeding a preset threshold value in the text information through the positioning model after the text information is input into a pre-trained positioning model for carrying out fuzzy identification on the characters, so that the characters are used as target characters;

The determining unit is used for searching the geographic position information containing the target characters from the corresponding relation between the characters and the geographic positions for each target character so as to determine the geographic positions in the geographic position information as the geographic positions corresponding to the target characters.

In one possible embodiment, the apparatus further comprises:

the system comprises a sample unit, a training text and a processing unit, wherein the sample unit is used for converting the sample voice data into training texts through a voice recognition technology for each sample voice data after a plurality of sample voice data are acquired;

the marking unit is used for marking a first mark for a first character used for representing the name of the place name in the training text and marking a second mark for a non-first character used for representing the name of the place name, so that the training text carrying the first mark and the second mark is used as a target training text;

an input unit for inputting a dataset comprising a plurality of target training texts into a bert+crf model for training the bert+crf model into the positioning model by means of supervised learning.

In a possible embodiment, the input unit is specifically configured to:

the alignment unit is used for placing the geographic position, at least one word or word used for representing the geographic position name and the corresponding relation between the at least one word or word and the geographic position into the geographic position set corresponding to the geographic position from at least one geographic position information corresponding to the geographic position according to each geographic position in the geographic position information after acquiring at least one geographic position information comprising the geographic position, the word used for representing the geographic position name and the corresponding relation between the word and the geographic position;

the first storage unit is used for storing the geographical position set into the corresponding relation between the characters and the geographical positions.

In one possible embodiment, the apparatus further comprises:

the device comprises a presetting unit, a processing unit and a processing unit, wherein the presetting unit is used for taking a marked vocabulary as a preset place name for each marked vocabulary before acquiring at least one geographic position information comprising a geographic position, a character used for representing the name of the geographic position and the corresponding relation between the character and the geographic position, and setting the geographic position corresponding to the preset place name for each preset place name; the mark vocabulary is a vocabulary formed by characters carrying target marks, and the target marks comprise a first mark and a second mark;

and the second storage unit is used for storing the preset place names, the geographic positions of the preset place names and the corresponding relation between the preset place names and the geographic positions into geographic position information corresponding to the preset place names according to each preset place name.

In a third aspect, an embodiment of the present application further provides an electronic device, including: a processor, a storage medium storing machine-readable instructions executable by the processor, the processor and the storage medium communicating over a bus when the electronic device is running, the processor executing the machine-readable instructions to perform the steps of the method of any of the first aspects.

In a fourth aspect, embodiments of the present application also provide a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the method according to any of the first aspects.

After voice data is acquired, the voice data is converted into text information through a voice recognition technology; after converting voice data into text information, as the voice data is easily interfered by environment, noise, signals and the like, the text information converted by the voice recognition technology has certain errors, and partial characters in the text information can be changed into other characters by the errors; for each target text, searching the geographic position information containing the target text from the corresponding relation between the text and the geographic position, so as to determine the geographic position in the geographic position information as the geographic position corresponding to the target text. Under the condition that errors are easy to occur in voice conversion texts, compared with the method for determining the place names by word-by-word or word-by-word comparison in the prior art, the method can identify the words with the semantic similarity exceeding the preset threshold value in voice data through the positioning model, namely, when wrongly written characters exist in the converted place names, the positioning model can also identify the place names, accuracy rate of recognizing the place names in the voice is improved, and geographic positions corresponding to the place names are determined for the place names.

In order to make the above objects, features and advantages of the present application more comprehensible, preferred embodiments accompanied with figures are described in detail below.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings that are needed in the embodiments will be briefly described below, it being understood that the following drawings only illustrate some embodiments of the present application and therefore should not be considered limiting the scope, and that other related drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 shows a flowchart of a location identification method according to an embodiment of the present application.

Fig. 2 shows a flowchart of another location identification method provided in an embodiment of the present application.

Fig. 3 shows a schematic structural diagram of a position identifying device according to an embodiment of the present application.

Fig. 4 shows a schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed Description

For the purpose of making the objects, technical solutions and advantages of the embodiments of the present application more clear, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application, and it should be understood that the accompanying drawings in the present application are only for the purpose of illustration and description, and are not intended to limit the protection scope of the present application. In addition, it should be understood that the schematic drawings are not drawn to scale. A flowchart, as used in this application, illustrates operations implemented according to some embodiments of the present application. It should be understood that the operations of the flow diagrams may be implemented out of order and that steps without logical context may be performed in reverse order or concurrently. Moreover, one or more other operations may be added to the flow diagrams and one or more operations may be removed from the flow diagrams as directed by those skilled in the art.

In addition, the described embodiments are only some, but not all, of the embodiments of the present application. The components of the embodiments of the present application, which are generally described and illustrated in the figures herein, may be arranged and designed in a wide variety of different configurations. Thus, the following detailed description of the embodiments of the present application, as provided in the accompanying drawings, is not intended to limit the scope of the application, as claimed, but is merely representative of selected embodiments of the application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present application without making any inventive effort, are intended to be within the scope of the present application.

It should be noted in advance that the term "comprising" will be used in this application to indicate the presence of the features stated hereinafter, but not to preclude the addition of further features.

It should be noted in advance that, the apparatus, the electronic device, or the like according to the embodiments of the present application may be executed on a single server, or may be executed on a server group. The server group may be centralized or distributed. In some embodiments, the server may be local or remote to the terminal. For example, the server may access information and/or data stored in a service requester terminal, a service provider terminal, or a database, or any combination thereof, via a network. As another example, the server may be directly connected to at least one of the service requester terminal, the service provider terminal, and the database to access the stored information and/or data. In some embodiments, the server may be implemented on a cloud platform; for example only, the cloud platform may include a private cloud, public cloud, hybrid cloud, community cloud (community cloud), distributed cloud, inter-cloud (inter-cloud), multi-cloud (multi-cloud), and the like, or any combination thereof.

In addition, the apparatus or the electronic device according to the embodiments of the present application may be executed on an access device or a third party device, and may specifically include: a mobile device, a tablet computer, a laptop computer, or a built-in device in a motor vehicle, etc., or any combination thereof. In some embodiments, the mobile device may include a smart home device, a wearable device, a smart mobile device, a virtual reality device, or an augmented reality device, or the like, or any combination thereof. In some embodiments, the smart home device may include a control device for a smart appliance device, a smart monitoring device, a smart television, a smart camera, or an intercom, or the like, or any combination thereof. In some embodiments, the wearable device may include a smart bracelet, a smart helmet, a smart watch, a smart accessory, etc., or any combination thereof. In some embodiments, the smart mobile device may include a smart phone, a personal digital assistant (Personal Digital Assistant, PDA), a gaming device, a navigation device, etc., or any combination thereof. In some embodiments, the virtual reality device and/or the augmented reality device may include a virtual reality helmet, an augmented reality helmet, or the like, or any combination thereof. For example, the virtual reality device and/or the augmented reality device may include various virtual reality products, and the like.

Example 1

Fig. 1 is a flowchart of a position identifying method according to an embodiment of the present application, as shown in fig. 1, where the method is implemented by the following steps:

step 101, after voice data is acquired, the voice data is converted into text information through a voice recognition technology.

Specifically, the voice data is data including voice of the user, and may be, for example, a piece of voice information sent by the user, call data of the user, and the like. The application scene of the embodiment of the application is an insurance industry accident report and claim settlement scene. Therefore, in the embodiment of the present application, the voice data is call data between the user and the intelligent robot when service feedback or fault reporting is performed. After the voice data is acquired, the voice data is converted into text information by a voice recognition technology, and the text information is cleaned. By cleaning the text information, eliminating the disturbing information such as messy codes, special characters and the like generated in the text information during the voice conversion process, and deleting the blank text.

Step 102, after inputting the text information into a pre-trained positioning model for carrying out fuzzy recognition on characters, recognizing the characters with the semantic similarity exceeding a preset threshold value with the place names in the text information through the positioning model, and taking the characters as target characters.

Specifically, the positioning model can identify the characters in the text information, and the characters with the semantic similarity exceeding a preset threshold value with the place name in the text information are identified according to parameters obtained by training the positioning model in advance. The target text may be one word or a vocabulary consisting of at least two words. After step 101 is executed, text information after cleaning is obtained, the text information is input into a positioning model, the positioning model identifies characters in the text information, marks the identified characters, and extracts the characters marked with the marks to obtain target characters, wherein the semantic similarity of the target characters and place names in the text information exceeds a preset threshold value. The parameter of the positioning model determines the semantic similarity of the extracted target characters and the place names, and the more accurate the parameter obtained by training the positioning model is, the higher the semantic similarity of the extracted target characters and the place names is.

In practical applications, when the quality of a voice call is affected by the problems of the user such as the conventional word, speaking mode, accent or call environment, call noise, call signal, etc., some errors may occur when the voice is converted into text by the voice recognition technology, for example, "i am to the day" is converted into "i am to the day" and "i am to the day" are. Therefore, through the trained positioning model, words similar to the semantic meaning of "Tianjing" or "Tianjing" in the voice call data can be identified as target characters, and the specific training process of the positioning model is seen in step 211-step 215.

Step 103, for each target text, searching the geographic position information containing the target text from the corresponding relation between the text and the geographic position, so as to determine the geographic position in the geographic position information as the geographic position corresponding to the target text.

Specifically, the corresponding relation between the characters and the geographic positions comprises a knowledge graph constructed for at least one geographic position; for each geographic location, the knowledge graph includes entities, geographic locations, relationships between entities and geographic locations, and relationships between different entities. The knowledge graph is essentially a knowledge base of a semantic network and is a relationship graph used for representing the relationship between different entities; wherein the relationship is used to express a relationship between different entities; an entity is a noun used to represent an object or law of existence in the real world, such as a name of a person, a place, a concept, a medicine, a company, etc.

The corresponding relation between the words and the geographic positions comprises a plurality of pieces of geographic position information or a plurality of pieces of geographic position sets, and each piece of geographic position information or each piece of geographic position set comprises words representing the names of the geographic position information, the geographic positions of the geographic position information and the corresponding relation between the words and the geographic positions. After the positioning model extracts the target text in step 102, for each target text, the geographic position information including the target text is searched from the corresponding relation between the text and the geographic position, and the geographic position in the geographic position information is determined as the geographic position corresponding to the target text.

When only one target character is identified in one text message, a geographic position can be uniquely determined according to the target character; when a plurality of target characters are identified in one text message, determining the geographic position corresponding to the text message according to the geographic positions determined for the plurality of target characters from the corresponding relation between the characters and the geographic positions.

The rule for determining the geographic position corresponding to the text information can be preset, the set rule is not limited, and the rule can be adjusted according to actual conditions. For example, determining the geographic position with the smallest range corresponding to the plurality of target characters as the geographic position corresponding to the text information, wherein the characters in the text information are "Beijing going plaza A watching flag raising", the target characters which can be identified by the positioning model are "Beijing", "plaza A", and the geographic position corresponding to the text information is determined as the geographic position of "plaza A".

After determining the geographic position of the target text, the target text and the geographic position corresponding to the target text can be displayed through a webpage or a client, the target text and the geographic position can be marked to the voice information, or the data can be directly stored, marked and output to a third party.

For example, the following geographical location information is stored in the correspondence between the text and the geographical location:

information one, beijing, north latitude 39°26'-41°03', east longitude 115°25'-117°30'.

Information II, shanghai, north latitude 30 DEG 40'-31 DEG 53', east longitude 120 DEG 52'-122 DEG 12'.

Assuming that the target text identified by the positioning model is "Shanghai", searching the target text "Shanghai" in the geographic position information in the corresponding relation between the text and the geographic position, determining that the second information is the geographic position information containing the target text "Shanghai", and determining the geographic position "North latitude 30 degrees 40 degrees '-31 degrees 53', east longitude 120 degrees 52 degrees '-122 degrees 12'" in the second information as the geographic position corresponding to the target text "Shanghai".

In a possible implementation manner, fig. 2 is a flowchart of another location identification method provided in the example of the present application, and as shown in fig. 2, the method is implemented by the following steps:

step 201, after a plurality of sample voice data are acquired, for each sample voice data, converting the sample voice data into training text through a voice recognition technology.

Specifically, in the embodiment of the present application, the sample voice data is user history voice data including call data of the user and the intelligent robot, which is obtained from a server or a cloud. After a plurality of sample voice data are obtained, the sample voice data are converted into training texts through a voice recognition technology, the training texts are cleaned, interference information such as messy codes and special characters generated in the training texts in the voice conversion process is removed, and the blank texts are deleted.

And 202, marking a first identifier for a first character used for representing the name of the place name in the training text, and marking a second identifier for a second character which is not used for representing the name of the place name, so that the training text carrying the first identifier and the second identifier is used as a target training text.

Specifically, after a plurality of training texts are obtained in step 201, a first text in all place names in the training texts is marked with a first identifier, and the other text except the first text in all place names in the training texts is marked with a second identifier. In addition, the characters irrelevant to the place name are marked by a third mark. At this time, all characters in the training text are marked with the first identifier, the second identifier or the third identifier, and the training text carrying the first identifier, the second identifier and the third identifier is used as a target training text.

For example, if the text in the training text is "Beijing me is going to Beijing", the first identifier and the second identifier are marked by "Beijing" representing the place name, wherein the first word "north" of "Beijing" marks the first identifier; marking a second mark for a non-first word Beijing in Beijing; the third marks are marked for other words which are irrelevant to the place name in the word "me goes to Beijing in tonight" respectively.

Or, the text in the training text is "going to the back and hearing sound", and the text is "going to the back" in "going to the back" and marking the first mark; marking the second mark for both in both; a third identifier is marked for each word in "go", "audible" respectively.

Step 203, inputting a data set containing a plurality of target training texts into a bert+crf model, so as to train the bert+crf model into the positioning model in a supervised learning manner.

Specifically, the BERT model (Bidirectional Encoder Representation from Transformers, a pre-training model) is used for increasing the generalization capability of the word vector model and fully describing character-level, word-level, sentence-level and even inter-sentence relationship features. CRF (Conditional Random Field ) for modeling the target sequence based on the observed sequence, focusing on solving the problem of serialization annotation. Supervised learning is a machine learning task that extrapolates functions from a labeled training dataset. In the embodiment of the application, a bert+crf model under a named entity recognition model is adopted, and the named entity recognition model can be changed according to actual conditions, for example, an entity recognition model such as Electra, ernie can be adopted. The named entity recognition process of the BERT+CRF model is a process of predicting a labeling sequence according to an input sentence, rules can be summarized from training texts, and unknown texts can be predicted according to the rules.

After the target training text is marked with the first identifier, the second identifier and the third identifier in step 202, a dataset composed of a plurality of target training texts is input into the bert+crf model.

In one possible embodiment, in performing step 203, this is achieved specifically by:

step 211, respectively placing the target training files in the data set into a training set, a verification set and a test set according to a preset proportion.

Specifically, in the embodiment of the present application, the preset ratio is 7:2:1, that is, 70% of the target training texts in the data set are placed in the training set, 20% of the target training texts in the data set are placed in the verification set, and 10% of the target training texts are placed in the test set. The embodiment of the application does not limit the preset proportion, and the preset proportion can be set according to the rest proportion.

Step 212, after the training set, the verification set and the test set are respectively input into the bert+crf model, training the bert+crf model under the super parameters through at least one target training text in the training set and the first identifier and the second identifier carried by the at least one target training text for each super parameter preset for the bert+crf model to obtain parameters of the bert+crf model under the super parameters; and the BERT+CRF model identifies characters with semantic similarity exceeding a preset threshold value with the place name in the text information according to the parameters.

Specifically, the parameters are variables that the model can automatically learn from data, such as weights, deviations, etc. for deep learning. Super-parameters are parameters used to determine a model, which are different for the same model, under different super-parameters, which are typically empirically determined variables. For example, in deep learning, the super parameters are: learning rate, number of iterations, number of layers, number of neurons per layer, etc. The preset threshold is preset, and can be set to be 100%, 90% or the like according to practical situations.

And respectively inputting a training set, a verification set and a test set into the BERT+CRF model, training the model through the training set, verifying the trained model through the verification set, and testing the verified model through the test set. Before the BERT+CRF model is trained, a plurality of super parameters are set for the model, and the parameters of the BERT+CRF model under the super parameters are trained under each super parameter through a training set. After the training set is input into the BERT+CRF model, characters carrying the first mark and the second mark in the target training text of the training set are used as expected output values, the target training text in the training set is analyzed through the BERT+CRF model to obtain actual output values, and the parameters of the BERT+CRF model under the super parameters are determined by continuously adjusting errors between the actual output values and the expected output values, so that the BERT+CRF model can generate the ability of predicting unknown samples according to the parameters. For example, when the number of iterations (superparameters) set by the same model is different, the parameters obtained by training the model through the training set are also different under different iteration numbers, and the different parameters enable the model to be different in the ability of identifying characters with semantic similarity with place names exceeding a preset threshold under different superparameters. The more accurate the super parameter setting is, the higher the semantic similarity between the characters and place names of the parameter identification obtained by training the model under the super parameter is.

Step 213, after obtaining the parameters, verifying the parameters of the bert+crf model under the super parameters by the verification set for each super parameter, so as to obtain a first recognition rate of the model under the parameters; the first recognition rate is the success rate of recognizing characters with semantic similarity with place names exceeding a preset threshold value in the target training texts in the verification set by the BERT+CRF model under the parameter.

Specifically, after step 212 is performed, parameters corresponding to the model under different super parameters are obtained. Inputting the verification set into the BERT+CRF model under the super parameters aiming at the corresponding BERT+CRF model under each super parameter, identifying characters in the target training text of the verification set according to the parameters of the BERT+CRF model under the super parameters, and calculating a first identification rate by comparing the identification result with an identification result of an identification carried by the target training text. The recognition result refers to the recognition condition of the BERT+CRF model on the characters in the target training text of the verification set under the parameter. The first recognition rate is the ratio of the number of place names successfully recognized to the total number of place names in the target training text.

For example, when text information in one of the target training texts in the verification set is "sit up to play Beijing today", the three marks are marked by seven words of "sit up to play" and "play" of the target training text in the verification set in advance respectively, the first mark is marked by "north", and the second mark is marked by "Beijing";

for a model under one of the super parameters, inputting text information ' today's high-speed iron to Beijing playing ' in a target training text into the trained model, and marking the text information with the parameters obtained through training, namely a first verification identifier, a second verification identifier and a third verification identifier for representing a first identifier, a second identifier and a third identifier, respectively, if the model marks the third verification identifier for ' today's high-speed iron to ' playing ' seven words according to the parameters, marking the first verification identifier for ' north ', and marking the second verification identifier for ' Beijing '. And comparing the vocabulary formed by the first mark and the second mark, wherein the model can identify the place name of Beijing, the obtained first identification rate is 100% for the target training text, and when the target training text in the verification set is a plurality of target training texts, the first identification rate of the model for the whole verification set is calculated.

Step 214, after determining the super-parameters of the bert+crf model according to the first recognition rate, testing the parameters of the bert+crf model under the super-parameters through a test set to obtain a second recognition rate of the model under the parameters; the second recognition rate is the success rate of recognizing characters with semantic similarity with place names exceeding a preset threshold value in the target training texts in the test set by the BERT+CRF model under the parameter.

Specifically, after step 213 is performed, a plurality of first recognition rates of the model are obtained, and the hyper-parameters of the model corresponding to the highest first recognition rate are determined as the hyper-parameters of the bert+crf model; or determining the super-parameters of the model corresponding to the first recognition rate exceeding the threshold value as the super-parameters of the BERT+CRF model; the method for determining the super parameter is not limited in the embodiment of the application. And after the super-parameters are determined for the BERT+CRF model, taking the target training files in the test set as test samples, and testing the parameters of the BERT+CRF model under the super-parameters through the test set. The second recognition rate is the same as the first recognition rate.

And step 215, judging whether the second recognition rate is greater than or equal to a preset recognition rate, and when the second recognition rate of the BERT+CRF model is greater than or equal to the preset recognition rate, taking the model as the positioning model to recognize characters with semantic similarity with place names exceeding a preset threshold value in the text information through the positioning model.

Specifically, after step 214, a second recognition rate obtained by testing the model has been obtained, and it is determined whether the second recognition rate is greater than a preset recognition rate. When the second recognition rate is smaller than the preset recognition rate, the recognition effect of the positioning model under the super parameter obtained through training on the place name is not ideal, and the super parameter of the model needs to be readjusted and the positioning model under the super parameter is trained, or the BERT+CRF model is replaced by other named entity recognition models for training. When the second recognition rate is larger than the preset recognition rate, the model with the second recognition rate is considered to be capable of achieving recognition accuracy of the place names in the text, and the model can be used as a positioning model, so that characters with semantic similarity exceeding a preset threshold value with the place names in the text information are recognized through the positioning model.

In a possible implementation, the correspondence between the words and the geographic location is constructed by the following steps:

after at least one geographic position information containing a geographic position, a word used for representing the name of the geographic position and the corresponding relation between the word and the geographic position is obtained, for each geographic position in the geographic position information, the geographic position, at least one word or vocabulary used for representing the name of the geographic position and the corresponding relation between the at least one word or vocabulary and the geographic position are put into a geographic position set corresponding to the geographic position from at least one geographic position information corresponding to the geographic position; and storing the geographical position set into the corresponding relation between the characters and the geographical positions.

Specifically, the geographic position information can be obtained through a network crawling method, a historical service data sorting method and the like. The historical service data comprises geographic position information comprising geographic positions, characters used for representing names of the geographic positions and corresponding relations between the characters and the geographic positions. The method comprises the steps of performing operations such as entity disambiguation, entity alignment, attribute alignment and the like on geographic position information, and aims to sort geographic position information with the same geographic position and different place names into the same geographic position set. In the geographic position set, each place name uniquely corresponds to one geographic position, and each geographic position corresponds to at least one place name.

For example, assume that there are a plurality of geographic location information:

information one, shandong, north latitude 34 deg. 22.9'-38 deg. 24.01', east longitude 114 deg. 47.5'-122 deg. 42.3'.

Information two, qilu, north latitude 34 DEG 22.9'-38 DEG 24.01', east longitude 114 DEG 47.5'-122 DEG 42.3'.

Information three, beijing, north latitude 39°26'-41°03', east longitude 115°25'-117°30'.

Four, imperial, north latitude 39 deg. 26'-41 deg. 03', east longitude 115 deg. 25'-117 deg. 30'.

Information five, beijing, north latitude 39 deg. 26'-41 deg. 03', east longitude 115 deg. 25'-117 deg. 30'.

The consolidated set of geographic information is:

integration (Shandong, qilu; north latitude 34 DEG 22.9'-38 DEG 24.01', east longitude 114 DEG 47.5 '-122 DEG 42.3').

Second set, (Beijing, didu, beijing; 39 DEG 26'-41 DEG 03', east meridian 115 DEG 25'-117 DEG 30').

The corresponding relation between the words and the geographic position also comprises the words and the geographic position obtained from the historical service data, the network and other channels and the relation between the words and the geographic position.

For example, "Zhang Sanzhu resides in Beijing", where "Zhang Sanj" and "Beijing" are entities in the relationship map, in the embodiment of the present application, "Beijing" is an entity for representing a geographic location, corresponds to a geographic location in the knowledge map representing the latitude and longitude of Beijing, and "Zhang Sanj" is an entity related to "Beijing" which is an entity for representing a geographic location, and "Zhang Sanj" and "Beijing" are related to "residing". Then when "Beijing" is identified, it may be associated to "Zhang Sanj" according to "Beijing"; when "Zhang Sano" is identified, it may be associated to "Beijing" according to "Zhang Sano". So as to improve the accuracy of the determined geographic location based on the association between the words.

For example, after the knowledge graph includes "Zhang san", "Beijing" and the relationship between "Zhang san" and "Beijing", when the text information includes the information with mispronounced words such as "Zhang san resides in Bei Jing", the geographic location can be inferred from the corresponding relationship between the words and the geographic location as the geographic location where Beijing is located.

In the embodiment of the application, the knowledge graph is mainly used for representing the relationship between entities of the geographic location.

For example: according to the historical service information or the information obtained by network acquisition, the following text is arranged:

text one, go to square a in beijing.

Text two, square B, is next to square a.

The "Beijing", "Square A" and "Square B" in the above-mentioned text all belong to the entity used for expressing the geographic location, the relation among every entity is, "square A" is "Beijing", "square B" and "square A" are adjacent, produce the relation through "square A" between "Beijing" and "square B", store the said relation in the correspondent relation of the said characters and geographic location.

In a possible embodiment, before obtaining at least one geographic location information including a geographic location, a text for representing a name of the geographic location, and a correspondence between the text and the geographic location, the following steps are further performed:

For each marked word, taking the marked word as a preset place name, and setting a geographic position corresponding to the preset place name for each preset place name; the marked vocabulary is a vocabulary formed by words carrying target marks, and the target marks comprise a first mark and a second mark. And storing the preset place names, the geographic positions of the preset place names and the corresponding relation between the preset place names and the geographic positions into geographic position information corresponding to the preset place names aiming at each preset place name.

Specifically, in executing step 202, for each word in the training text that is used to represent the place name, the first identifier or the second identifier is marked. And regarding each word carrying the first mark and at least one word carrying the second mark, taking the word or the word as a preset place name, setting a geographic position corresponding to the preset place name for each preset place name, and setting geographic position information corresponding to the preset place name according to the preset place name and the geographic position.

For example, the target training text is marked with characters such as "beijing", "dorjing", "Tianjin", "Tianjing", "lu" and "shandong", respectively. Marking the first mark for north, back, sky, robust and mountain; marking the second marks for 'Beijing', 'Jing', 'jin', 'Jing', 'Dong'; taking the text 'robust' carrying the first mark as a first preset place name; the words "Beijing", "Tianjin", "Tianjing" and "Shandong" formed by the words carrying the first and second marks are used as the second, third, fourth, fifth and sixth preset place names. And setting corresponding geographic positions for the six preset place names respectively. And for each preset place name, storing the preset place name, the geographic position of the preset place name and the corresponding relation between the preset place name and the geographic position into geographic position information corresponding to the preset place name, and setting six pieces of geographic position information corresponding to six preset place names:

Information one, robust, north latitude 34 ° 22.9'-38 ° 24.01', east longitude 114 ° 47.5'-122 ° 42.3'.

Information two, beijing, north latitude 39 deg. 26'-41 deg. 03', east longitude 115 deg. 25'-117 deg. 30'.

Information three, beijing, north latitude 39 ° -41 ° -03 ', east longitude 115 ° -25 ° -117 ° 30'.

Information four, tianjin, north latitude 38 deg. 34'-40 deg. 15', east longitude 116 deg. 43'-118 deg. 04'.

Five information, tokyo, north latitude 38 deg. 34'-40 deg. 15', east longitude 116 deg. 43'-118 deg. 04'.

Information six, shandong, north latitude 34 deg. 22.9'-38 deg. 24.01', east longitude 114 deg. 47.5'-122 deg. 42.3'.

Example two

Fig. 3 is a schematic structural diagram of a position identifying device according to an embodiment of the present application, as shown in fig. 3, where the device includes: a conversion unit 301, a positioning unit 302, a determination unit 303.

The converting unit 301 is configured to convert the voice data into text information by a voice recognition technology after the voice data is acquired.

And the positioning unit 302 is used for identifying the characters with the semantic similarity with the place names in the text information exceeding a preset threshold value through the positioning model after inputting the text information into a pre-trained positioning model for carrying out fuzzy identification on the characters, so as to take the characters as target characters.

And a determining unit 303, configured to, for each target text, search, from the corresponding relationship between the text and the geographic location, geographic location information including the target text, so as to determine the geographic location in the geographic location information as the geographic location corresponding to the target text.

In one possible embodiment, the apparatus further comprises:

and the sample unit is used for converting the sample voice data into training texts through a voice recognition technology for each sample voice data after a plurality of sample voice data are acquired.

The marking unit is used for marking a first mark for the first character used for representing the name of the place name in the training text, and marking a second mark for the non-first character used for representing the name of the place name, so that the training text carrying the first mark and the second mark is used as a target training text.

In a possible embodiment, the input unit is specifically configured to:

and respectively putting the target training files in the data set into a training set, a verification set and a test set according to a preset proportion.

After the training set, the verification set and the test set are respectively input into the BERT+CRF model, training the BERT+CRF model under the super parameters through at least one target training text in the training set and the first identifier and the second identifier carried by the at least one target training text aiming at each super parameter preset for the BERT+CRF model so as to obtain parameters of the BERT+CRF model under the super parameters; and the BERT+CRF model identifies characters with semantic similarity exceeding a preset threshold value with the place name in the text information according to the parameters.

After obtaining the parameters, verifying the parameters of the BERT+CRF model under the super parameters through the verification set for each super parameter to obtain a first recognition rate of the model under the parameters; the first recognition rate is the success rate of recognizing characters with semantic similarity with place names exceeding a preset threshold value in the target training texts in the verification set by the BERT+CRF model under the parameter.

After the super-parameters of the BERT+CRF model are determined according to the first recognition rate, testing the parameters of the BERT+CRF model under the super-parameters through a testing set to obtain a second recognition rate of the model under the parameters; the second recognition rate is the success rate of recognizing characters with semantic similarity with place names exceeding a preset threshold value in the target training texts in the test set by the BERT+CRF model under the parameter.

the alignment unit is used for placing the geographic position, at least one word or word used for representing the geographic position name and the corresponding relation between the at least one word or word and the geographic position into the geographic position set corresponding to the geographic position from at least one geographic position information corresponding to the geographic position according to each geographic position in the geographic position information after at least one geographic position information containing the geographic position, the word used for representing the geographic position name and the corresponding relation between the word and the geographic position is obtained.

In one possible embodiment, the apparatus further comprises:

the device comprises a presetting unit, a processing unit and a processing unit, wherein the presetting unit is used for taking a marked vocabulary as a preset place name for each marked vocabulary before acquiring at least one geographic position information comprising a geographic position, a character used for representing the name of the geographic position and the corresponding relation between the character and the geographic position, and setting the geographic position corresponding to the preset place name for each preset place name; the marked vocabulary is a vocabulary formed by words carrying target marks, and the target marks comprise a first mark and a second mark.

Example III

Fig. 4 is a schematic structural diagram of an electronic device according to a third embodiment of the present application, including: a processor 401, a storage medium 402 and a bus 403, the storage medium 402 storing machine-readable instructions executable by the processor 401, when the electronic device runs the method as in the first embodiment, the processor 401 communicates with the storage medium 402 through the bus 403, and the processor 401 executes the machine-readable instructions to perform the steps as in the first embodiment.

In the embodiment of the present application, the storage medium 402 may further execute other machine readable instructions to perform the method as described in the other embodiments, and the specific implementation of the steps and principles of the method are referred to in the description of the embodiment, and are not described in detail herein.

Example IV

The fourth embodiment of the present application also provides a computer readable storage medium having a computer program stored thereon, the computer program being executed by a processor when executed to perform the steps as in the first embodiment.

In the embodiment of the present application, the computer program may further execute other machine readable instructions when executed by the processor to perform the method as described in the other embodiments, and the specific implementation of the steps and principles of the method are referred to in the description of the embodiment, and are not described in detail herein.

In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods may be implemented in other manners. The above-described apparatus embodiments are merely illustrative, and the division of the modules is merely a logical function division, and there may be additional divisions when actually implemented, and for example, multiple modules or components may be combined or integrated into another system, or some features may be omitted or not performed. Alternatively, the coupling or direct coupling or communication connection shown or discussed with each other may be through some communication interface, indirect coupling or communication connection of devices or modules, electrical, mechanical, or other form.

The modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical units, may be located in one place, or may be distributed over multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

In addition, each functional unit in each embodiment of the present application may be integrated in one processing unit, or each unit may exist alone physically, or two or more units may be integrated in one unit.

The functions, if implemented in the form of software functional units and sold or used as a stand-alone product, may be stored in a non-volatile computer readable storage medium executable by a processor. Based on such understanding, the technical solution of the present application may be embodied essentially or in a part contributing to the prior art or in a part of the technical solution, in the form of a software product stored in a storage medium, including several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the embodiments of the present application. And the aforementioned storage medium includes: a usb disk, a removable hard disk, a ROM, a RAM, a magnetic disk, or an optical disk, etc.

The foregoing is merely a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think about changes or substitutions within the technical scope of the present application, and the changes or substitutions are covered in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

1. A method of location identification, comprising:

for each target character, searching geographical position information containing the target character from the corresponding relation between the character and the geographical position, and determining the geographical position in the geographical position information as the geographical position corresponding to the target character;

the positioning model is trained by:

Inputting a dataset comprising a plurality of target training texts into a bert+crf model to train the bert+crf model into the positioning model by means of supervised learning;

inputting a dataset comprising a plurality of target training texts into a bert+crf model to train the bert+crf model into the positioning model by means of supervised learning, comprising:

2. The method of claim 1, wherein the text to geographic location correspondence is constructed by:

3. The method of claim 2, wherein prior to obtaining at least one geographic location information comprising a geographic location, text representing a name of the geographic location, and a correspondence of the text to the geographic location, the method further comprises:

4. A position identification apparatus, the apparatus comprising:

The determining unit is used for searching geographic position information containing the target characters from the corresponding relation between the characters and the geographic positions for each target character so as to determine the geographic positions in the geographic position information as the geographic positions corresponding to the target characters;

the apparatus further comprises:

an input unit for inputting a dataset comprising a plurality of target training texts into a bert+crf model to train the bert+crf model into the positioning model by means of supervised learning;

the input unit is specifically configured to:

5. An electronic device, comprising: a processor, a storage medium and a bus, the storage medium storing machine-readable instructions executable by the processor, the processor and the storage medium communicating over the bus when the electronic device is running, the processor executing the machine-readable instructions to perform the steps of the location identification method of any of claims 1 to 3.

6. A computer-readable storage medium, characterized in that it has stored thereon a computer program which, when executed by a processor, performs the steps of the position identification method according to any of claims 1 to 3.