Data processing method, device, electronic device and computer storage medium

Abstract

Embodiments of the present invention provide a data processing method, apparatus, electronic device, and computer storage medium. Among them, the data processing method includes extracting features from the address information in the user's dialogue statement to obtain the address feature information of the dialogue statement; predicting an address word related to the address information according to the address feature information; generating an interaction statement for interacting with the user according to the address word, so as to determine the user's address information according to the dialogue statement of the user in response to the interaction statement. Through the embodiments of the present invention, accurate address information of the user can be obtained.

Description

Technical Field

[0001] Embodiments of the present invention relate to the field of computer technologies, and in particular, to a data processing method, apparatus, electronic device, and computer storage medium. Background Art

[0002] In security and customer service scenarios, the geographical location of a user is very important and determines the entry point for service provision. However, it has been found in practice that usually the address provided by the user is an informal short address information with ambiguity, making the address information obtained by the service system inaccurate and unable to provide accurate, timely, and good services for the user. Therefore, it is of great significance to be able to quickly obtain more address information and make the address information accurate. Summary of the Invention

[0003] In view of this, embodiments of the present invention provide a data processing solution to solve some or all of the above problems.

[0004] According to a first aspect of embodiments of the present invention, there is provided a data processing method, including: extracting features of address information in a user's dialogue statement to obtain address feature information of the dialogue statement; predicting an address word related to the address information according to the address feature information; generating an interaction statement for interacting with the user according to the address word, so as to determine the user's address information according to a dialogue statement of the user in response to the interaction statement.

[0005] According to a second aspect of embodiments of the present invention, there is provided a data processing apparatus, including a feature extraction module, configured to extract features of address information in a user's dialogue statement to obtain address feature information of the dialogue statement; a prediction module, configured to predict an address word related to the address information according to the address feature information; and an interaction module, configured to generate an interaction statement for interacting with the user according to the address word, so as to determine the user's address information according to a dialogue statement of the user in response to the interaction statement.

[0006] According to a third aspect of embodiments of the present invention, there is provided an electronic device, including: a processor, a memory, a communication interface, and a communication bus, where the processor, the memory, and the communication interface complete communication with each other through the communication bus; the memory is configured to store at least one executable instruction, and the executable instruction causes the processor to perform operations corresponding to the data processing method according to the first aspect.

[0007] According to a fourth aspect of embodiments of the present invention, there is provided a computer storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the data processing method according to the first aspect.

[0008] According to the data processing solution provided by the embodiments of the present invention, feature extraction is performed on the user's dialogue statement, and an address word is predicted based on the extracted address feature information, so as to generate an interaction statement using the address word, and then, through the interaction statement, one or more rounds of conversations are carried out with the user to gradually obtain more accurate address information. This is more in line with human interaction habits, and can obtain accurate address information even when the user's dialogue statement only includes a vague short address, solving the problem in the prior art that the user's location cannot be accurately determined using a short address. Brief Description of the Drawings

[0009] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments recorded in the embodiments of the present invention, and those of ordinary skill in the art can also obtain other drawings based on these drawings.

[0010] Figure 1a It is a flowchart of the steps of a data processing method according to Embodiment 1 of the present invention;

[0011] Figure 1b It is a schematic diagram of the interaction between a dialogue system and a client in a usage scenario according to Embodiment 1 of the present invention;

[0012] Figure 1c It is a schematic diagram of generating an interaction statement in a usage scenario according to Embodiment 1 of the present invention;

[0013] Figure 2a It is a flowchart of the steps of a data processing method according to Embodiment 2 of the present invention;

[0014] Figure 2b It is a schematic diagram of the interaction between a dialogue system and a user in a usage scenario according to Embodiment 2 of the present invention;

[0015] Figure 3 It is a structural block diagram of a data processing device according to Embodiment 3 of the present invention;

[0016] Figure 4 It is a schematic diagram of the structure of an electronic device according to Embodiment 4 of the present invention. Detailed Embodiments

[0017] To enable those skilled in the art to better understand the technical solutions in the embodiments of the present invention, the following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments in the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art shall fall within the scope protected by the embodiments of the present invention.

[0018] The following further illustrates the specific implementation of the embodiments of the present invention in conjunction with the accompanying drawings of the embodiments of the present invention.

[0019] Embodiment 1

[0020] Referring to Figure 1a , a step flowchart of a data processing method according to Embodiment 1 of the present invention is shown.

[0021] In the prior art, in scenarios such as public security and customer service, when providing services to users, users are required to provide complete, clear, and accurate address information, such as "No. 1, Pinghai Road, Shangcheng District, Hangzhou City, Zhejiang Province, Hangzhou Stomatological Hospital". For users, it is difficult to obtain complete address information, especially in relatively unfamiliar areas, where the difficulty will be even greater, and the time-consuming to obtain complete address information may be relatively long, which is not applicable in some emergency situations, resulting in the inability to provide services to users in a timely and good manner.

[0022] The data processing method in this embodiment can be used to solve the foregoing problems. This method can be executed on a terminal device or on a server (the server includes a server or the cloud). This embodiment takes execution in the cloud as an example for illustration.

[0023] The data processing method in this embodiment includes the following steps:

[0024] Step S102: Extract features from the address information in the user's dialogue statement to obtain the address feature information of the dialogue statement.

[0025] The user's dialogue statement can be a statement that replies to an interactive statement generated by a machine. The interactive statement is, for example: "Where are you", "Is it the Metropark Hotel", and so on. The dialogue statement is, for example: "I am at the intersection of Pinghai Road and Zhonghe Middle Road", "Opposite the Metropark Hotel", "Yes", and so on.

[0026] The address information in the dialogue statement is used to indicate the user's location. Usually, the user may describe the location by referring to the geographical location and the corresponding relationship between the geographical location and the user's current location, such as "I am at the intersection of Pinghai Road and Zhonghe Middle Road", "Opposite the Metropark Hotel", and so on.

[0027] In this embodiment, the address information includes address keywords and auxiliary words of the address keywords. The address keywords are used to represent the reference geographical location, such as "Pinghai Road", "Zhonghe Middle Road", "Regent Hotel", and so on. The address keywords can be one or more than one.

[0028] Its auxiliary words are used to represent the corresponding relationship between the user's current location and the reference geographical location, such as "intersection", "opposite", "south", "north", and so on.

[0029] Any appropriate method can be used to extract the features of the address information, as long as the address feature information including the indication information for indicating the corresponding relationship between the user's current location and the reference geographical location can be obtained.

[0030] For example, in a specific implementation, the address feature information includes the indication information for indicating the corresponding relationship between the user's current location and the reference geographical location represented by the address keywords. Subsequently, the user's location can be predicted according to the address feature information, and then an address word can be generated based on the predicted location, and an interaction statement for interacting with the user can be generated based on the address word, so as to confirm whether the prediction is accurate, and then obtain the accurate address information of the user.

[0031] In this way, by interacting with the user in the form of a dialogue, when the user's dialogue statement only contains a vague short address, the user can be guided to include more address information in the dialogue statement to determine the accurate address information of the user.

[0032] Step S104: Predict an address word related to the address information according to the address feature information.

[0033] In a specific implementation manner, the POI (point of interest) that meets the address feature information can be retrieved from the address information library, and the address word can be determined according to the attributes of these POIs.

[0034] The geographical information library is used to store addresses with geographical location information (such as the coordinates of a certain geographical location, the complete address, etc.). Through the geographical information library, a query can be made for a certain geographical coordinate, and the POIs and their attributes (such as the query frequency of each POI) related to the queried geographical coordinate can be returned. Then, the predicted address word can be determined according to the attributes of the queried POIs.

[0035] For example, the address feature information includes indication information for indicating that the user's current location is opposite the Weijing Hotel. Then, the coordinates of the "Weijing Hotel" (denoted as coordinate A) can be obtained through the geographic information database. Furthermore, based on coordinate A, the POIs located opposite it and the attributes of the POIs can be queried. Based on the attributes of the POIs, the predicted address word can be determined. For example, the name of the POI opposite (such as "Hangzhou Stomatological Hospital") is determined as the predicted address word.

[0036] Of course, in other embodiments, other methods can be used to determine the predicted address word, and this embodiment does not limit this.

[0037] Step S106: Generate an interaction statement for interacting with the user based on the address word, so as to determine the user's address information according to the dialogue statement in response to the interaction statement by the user.

[0038] Those skilled in the art can generate the interaction statement based on the address word in any appropriate manner. For example, using a trained natural language generation model, input the address word into the natural language generation model, and obtain the output interaction statement, which contains the address word.

[0039] Taking the address word as "Hangzhou Stomatological Hospital" as an example, the interaction statement can be "Is it Hangzhou Stomatological Hospital?", "Are you at Hangzhou Stomatological Hospital?", etc.

[0040] The following combines Figure 1b and Figure 1c , and the implementation process in a specific alarm scenario is described as follows:

[0041] The dialogue system configured on the server side includes a semantic address parsing module and a geographic information database. Among them, the semantic address parsing module is used to obtain address feature information according to the dialogue statement and send it to the geographic information database. The geographic information database is used to determine relevant POIs according to the geographic feature information. The semantic address parsing module is also used to determine the predicted address word according to the attributes of the relevant POIs, generate an interaction statement according to the predicted address word, and send it to the user, so as to obtain the next dialogue statement.

[0042] As Figure 1c shown, the specific process is as follows:

[0043] When the semantic address parsing module determines that the user accesses, it generates an interaction statement (denoted as interaction statement 1) and displays it on the user's terminal device (as shown in the terminal interface in Figure 1c ), such as "May I ask where you are talking about". It should be noted that the interaction statement can be displayed in text form or in audio form, etc.

[0044] The user inputs dialogue statement 1, such as "The intersection of Pinghai Road and Zhonghe Middle Road", in response to interaction statement 1. The semantic address parsing module obtains address feature information based on dialogue statement 1 (which includes, for example, the indication information shown in Figure 1c indicating that the user's current location is at the intersection of Pinghai Road and Zhonghe Middle Road), and sends the address feature information to the geographic information database. The geographic information database determines the coordinates of the intersection based on the address feature information, and then determines the relevant POIs and returns them to the semantic address parsing module. The semantic address parsing module predicts the most likely POI (such as the Metropole Hotel) based on the relevant POIs, determines the address term (i.e., the Metropole Hotel) according to its attributes, and then generates interaction statement 2 (such as "Is it the Metropole Hotel?") based on the address term and sends it to the user.

[0045] The user inputs dialogue statement 2, such as "Opposite the Metropole Hotel", in response to interaction statement 2. The semantic address parsing model obtains address feature information based on dialogue statement 2 (which includes the indication information indicating that the user's current location is opposite the Metropole Hotel) and sends it to the geographic information database. The geographic information database determines the coordinates of the Metropole Hotel based on the address feature information, and then determines the relevant POIs (i.e., the POIs opposite it) and returns them to the semantic address parsing module. The semantic address parsing module predicts the most likely POI (such as the Hangzhou Stomatological Hospital) based on the relevant POIs, determines the address term (i.e., the Hangzhou Stomatological Hospital) according to its attributes, and then generates interaction statement 3 (such as "Is it the Hangzhou Stomatological Hospital?") based on the address term and sends it to the user.

[0046] The user inputs dialogue statement 3, such as "Yes", in response to interaction statement 3. The semantic address parsing module determines that dialogue statement 3 is a statement for confirming the address term in the previous interaction statement. Then, based on the address term in the previous interaction statement (i.e., the Hangzhou Stomatological Hospital), the accurate address information of the user can be determined, such as "Hangzhou Stomatological Hospital, No. 1, Pinghai Road, Shangcheng District, Hangzhou, Zhejiang Province".

[0047] It should be noted that in the foregoing usage scenario, the dialogue system including the semantic address parsing module and the geographic information database is used as an example for illustration, but the structure of the dialogue system is not limited to this, and the operations performed by the semantic address parsing module and the geographic information database are not limited to this. For example, after obtaining the relevant POIs, the geographic information database can predict the address term according to the attributes of the relevant POIs and then return it to the semantic address parsing module.

[0048] In this embodiment, feature extraction is performed on the user's dialogue statement, and an address word is predicted based on the extracted address feature information, so as to generate an interaction statement using the address word, and then through the interaction statement, one or more rounds of conversations are carried out with the user to gradually obtain more accurate address information. This is more in line with human interaction habits, and can obtain accurate address information even when the user's dialogue statement only includes a vague short address, solving the problem in the prior art that the short address cannot be used to accurately determine the user's location.

[0049] The data processing method of this embodiment can be executed by any suitable electronic device with data processing capabilities, including but not limited to: servers, mobile terminals (such as tablet computers, mobile phones, etc.) and PC machines, etc.

[0050] Embodiment Two

[0051] Refer to Figure 2a , which shows a step flowchart of a data processing method according to Embodiment Two of the present invention.

[0052] In this embodiment, a dialogue system configured on the server side (the server side includes a server or the cloud) is used as the execution subject to illustrate the implementation process of the data processing method. Of course, in other embodiments, the dialogue system can also be configured on the user's terminal device.

[0053] The data processing method of this embodiment includes the aforementioned steps S102 to S106. In addition, in this embodiment, the method further includes:

[0054] Step S100a: Determine whether the user's dialogue statement is a dialogue statement for confirming the address word in the previous interaction statement of the user's dialogue statement.

[0055] In a specific implementation, a trained neural network model (such as an LSTM model, an RNN model, etc.) can be used to perform semantic recognition on the dialogue statement to determine whether the dialogue statement is used to confirm the address word in the previous interaction statement.

[0056] When it is to confirm the address word in the previous interaction statement, perform address structuring processing on the address word in the previous interaction statement of the dialogue statement to obtain the address structuring information corresponding to the address word, and use the address structuring information as the user's address information; conversely, when it is not to confirm the address word in the previous interaction statement, steps S102 to S106 can be executed.

[0057] For example, if the previous interaction statement is "Is it the Grand Metropark Hotel?", and the dialogue statement is "Opposite the Grand Metropark Hotel", and through semantic recognition, it is determined that the dialogue statement is not a confirmation of the previous interaction statement, it means that the predicted address word in the previous interaction statement is not accurate. At this time, an operation of extracting the address information features in the user's dialogue statement can be performed, that is, steps S102 to S106 are executed to predict a new address word based on the user's dialogue statement and generate a new interaction statement.

[0058] Another example is that the previous interaction statement is "Is it the Hangzhou Stomatological Hospital?", and the dialogue statements are "Yes", "Right", etc. Through semantic recognition, it is determined that the dialogue statement represents a confirmation of the previous interaction statement, that is, the predicted address word in the previous interaction statement is accurate. At this time, the address word in the previous interaction statement of the user's dialogue statement can be processed for address structuring to obtain the address structured information corresponding to the address word, and the address structured information is used as the user's address information.

[0059] A specific way to process the address word in the previous interaction statement of the user's dialogue statement for address structuring is as follows: perform address completion processing on the address word in the previous interaction statement of the user's dialogue statement to obtain complete address information; perform address structuring processing on the complete address information to obtain the address structured information corresponding to the address word, and use the address structured information as the user's address information.

[0060] For address completion processing, a trained neural network model with address standardization function (such as LSTM, RNN, etc.) can be used. Input the address word into this model, and the corresponding completed complete address information can be obtained. For example, when "Hangzhou Stomatological Hospital" is input, the completed complete address information is "No. 1, Pinghai Road, Shangcheng District, Hangzhou, Zhejiang Province, Hangzhou Stomatological Hospital".

[0061] In addition, for address structuring processing of the complete address information, a trained neural network model with address standardization function (such as LSTM, RNN, etc.) can also be used, or other algorithms capable of address structuring can be used. This embodiment does not limit this.

[0062] Through a neural network model with address standardization function, each address segment in the complete address information can be determined and the address segment can be labeled to obtain address structured information. For example, for the complete address information "No. 1, Pinghai Road, Shangcheng District, Hangzhou, Zhejiang Province, Hangzhou Stomatological Hospital", the labeling result of the address segment can be "(Province) Zhejiang Province (City) Hangzhou City (District) Shangcheng District (Street, building number) No. 1, Pinghai Road (Detailed address) Hangzhou Stomatological Hospital". Of course, the address standardization labeling methods of different neural network models may be different, as long as each address segment in the complete address information can be labeled.

[0063] The address structured information obtained through address structuring processing is used as the user's address information. For example, in an e-commerce scenario, based on the user's address information, the shipping address information of the product is determined; in a public security scenario, based on the user's address information, the target address information for reporting an alarm is determined, and so on.

[0064] Optionally, in the case where the user's dialogue statement indicates that the address term in the previous interaction statement is inaccurate, step S100b can be executed before performing step S102.

[0065] Step S100b: Perform named entity recognition on the user's dialogue statement to obtain the address keywords and auxiliary words of the address keywords in the user's dialogue statement.

[0066] In the case where the dialogue statement includes address keywords and auxiliary words, before performing feature extraction on the address information through step S102, step S100b can be executed to determine the address keywords and auxiliary words in the dialogue statement.

[0067] In a specific implementation, the address keywords and auxiliary words can be obtained by using the method of named entity recognition. For example, the dialogue statement is "the intersection of Pinghai Road and Zhonghe Middle Road". By performing named entity recognition on it through CRF (Conditional Random Field), the determined result is, for example, "Pinghai Road (place name) and Zhonghe Middle Road (place name) intersection (auxiliary word)". Thus, the address keywords can be determined as Pinghai Road and Zhonghe Middle Road, and the auxiliary word corresponding to the address keywords is intersection.

[0068] Of course, in other embodiments, other methods can be used to determine the address keywords and auxiliary words, and this embodiment does not limit this.

[0069] Named Entity Recognition (NER for short, also known as "Proper Name Recognition") refers to the recognition of entities with specific meanings in the text, mainly including personal names, place names, organization names, proper nouns, etc.

[0070] In the case of performing step S100b, step S102 can be implemented as: matching the address feature extraction template for the address information according to the auxiliary words included in the address information; and performing feature extraction on the address information according to the address feature extraction template to obtain the address feature information of the dialogue statement.

[0071] Among them, the address feature extraction template can be a preset template. For example, through big data statistics and analysis, possible address feature extraction templates are determined in advance. The address feature extraction template includes at least one of the following: "[*] and [*] intersection", "[*] opposite", "[*] south", "[*] north", etc., but not limited to this. Among them, [*] is used to indicate an address keyword, and "[*] and [*] intersection" indicates that the user's location is at the intersection of the reference geographical locations represented by two address keywords. "[*] opposite" is used to indicate that the user's location is opposite the reference geographical location represented by the address keyword.

[0072] When matching, if it is determined that the word in the address feature extraction template is the same as the auxiliary word, or the semantic similarity between the word in the address feature extraction template and the auxiliary word meets the similarity threshold, it is determined that the address feature extraction template matches the auxiliary word. For example, if the auxiliary word is "intersection", then the address feature extraction template "[*] and [*] intersection" matches it.

[0073] When there are multiple nested auxiliary words in the dialogue statement, corresponding address feature extraction templates can be determined for each auxiliary word. For example, the dialogue statement is "opposite the intersection of Pinghai Road and Zhonghe Middle Road", where the auxiliary words include "intersection" and "opposite". The address feature extraction template matched by "intersection" is "[*] and [*] intersection", and the address feature extraction template matched by "opposite" is "[*] opposite".

[0074] The address feature information extracted according to the address feature extraction template is: the indication information for indicating that the user's current location is at the intersection of Pinghai Road and Zhonghe Middle Road; or the indication information for indicating that the user's current location is opposite the Metropark Hotel, etc.

[0075] In this way, the address feature information of the address information in the dialogue statement can be accurately extracted, so that the subsequent address word with higher accuracy can be obtained according to the address feature information, and then the speed of obtaining accurate address information can be improved.

[0076] After obtaining the address feature information, in a specific implementation manner of step S104, step S104 includes the following sub-steps:

[0077] Sub-step S1041: According to the address feature information, determine the geographical location coordinates corresponding to the address feature information.

[0078] For example, if the address feature information includes the indication information for indicating that the user's current location is opposite the Metropark Hotel, the corresponding geographical location coordinates can be the coordinates of the Metropark Hotel (the coordinates can be longitude and latitude, or longitude, latitude and altitude, etc.).

[0079] For another example, if the address feature information includes indication information indicating that the user's current location is at the intersection of Pinghai Road and Zhonghe Middle Road, the corresponding geographical location coordinates can be the coordinates of the intersection of Pinghai Road and Zhonghe Middle Road (with the reference geographical location being the intersection) (these coordinates can be longitude and latitude, or longitude, latitude, and altitude, etc.).

[0080] Sub-step S1042: Determine a matching geographical location search strategy according to the address feature extraction template corresponding to the address feature information.

[0081] It should be noted that this sub-step is an optional step. When executing step S104, this sub-step can be executed or omitted.

[0082] When executing this sub-step, the geographical location search strategy can be a pre-set strategy, which corresponds to the address feature extraction template. For example, if the address feature extraction template is "[*] opposite", it can correspond to the first geographical location search strategy; if the address feature extraction template is "[*] and [*] intersection", it can correspond to the second geographical location search strategy, and so on.

[0083] In this way, different geographical location search strategies can be determined for different address feature extraction templates corresponding to address feature information. Each address location search strategy contains one or more search processes, and executing this strategy can obtain the required POI and its attributes.

[0084] For example, the processes included in the first geographical location search strategy are as follows:

[0085] Process 1: Obtain the orientation information of the road where the reference geographical location (i.e., the address keyword contained in the corresponding address feature information, such as the Metropole Hotel) is located. For example, the Metropole Hotel is located on Pinghai Road, and its orientation is east-west.

[0086] Process 3: Determine the coordinate change trend of the opposite side of the reference geographical location according to the orientation of the road where it is located. For example, if the Metropole Hotel is located on the south side of Pinghai Road, the coordinate change trend of its opposite side should be that the north latitude gradually increases.

[0087] Sub-step S1043: Search for a pre-set geographical location according to the geographical location coordinates corresponding to the address feature information to obtain at least one candidate geographical location.

[0088] The pre-set geographical location can be multiple POIs pre-stored in the geographical information database. The candidate geographical location can be a POI that meets the requirements of the geographical location coordinates corresponding to the address feature information.

[0089] For example, the geographic location coordinates (hereinafter referred to as geographic location coordinates A) corresponding to the address feature information are the coordinates of the Metropark Hotel, and the candidate geographic location obtained by the search may be the POI opposite the Metropark Hotel.

[0090] For example, the geographic location coordinates (hereinafter referred to as geographic location coordinates A) corresponding to the address feature information are the coordinates of an intersection, then the candidate geographic location obtained by search can be a POI within a certain range of the intersection (the certain range can be set as needed, such as a diameter of 1 meter, 20 meters, 50 meters, 100 meters, 500 meters, 1000 meters, etc.).

[0091] In a specific implementation, if sub-step S1042 is executed, sub-step S1043 can be implemented as: through a geographic location search strategy, according to the geographic location coordinates corresponding to the address feature information, a preset geographic location is searched to obtain at least one candidate geographic location.

[0092] For example, in one case, the address feature information includes indication information for indicating that the user's geographical location is located opposite to the reference geographical location indicated by the address keyword. Then, through the geographical location search strategy, according to the geographical location coordinates corresponding to the address feature information, the following process is performed to search for the preset geographical location:

[0093] Process A: Get the direction information of the road where the reference geographical location is located. For example, the Metropark Hotel is located on Pinghai Road, which is east-west.

[0094] Process B: Through the geographic location search strategy, according to the direction information, the geographic location coordinates corresponding to the address feature information and the indication information, the preset geographic location is searched to obtain at least one candidate geographic location located opposite the reference geographic location.

[0095] For example, according to the direction information, it is determined that the latitude of the coordinates of the geographical location opposite the Metropark Hotel should gradually increase, so the preset geographical location is searched to obtain POIs as candidate geographical locations, whose longitude is the same as that of the reference geographical location or the longitude difference is within a certain range (the range can be set as needed), whose latitude is greater than the latitude of the reference geographical location, and whose distance from the reference geographical location is less than a certain value.

[0096] For example, in another case, the address feature information includes indication information for indicating that the user's geographical location is located at the intersection of two reference geographical locations indicated by the address keyword. Then, through the geographical location search strategy, according to the geographical location coordinates corresponding to the address feature information, the following process is performed to search for the preset geographical location:

[0097] Process i: Through the geographical location search strategy, based on the geographical location coordinates and indication information corresponding to the address feature information, search the preset geographical locations to obtain at least one candidate geographical location of the intersection located at the reference geographical location.

[0098] For example, according to the reference geographical locations "Pinghai Road" and "Zhonghe Middle Road" indicated by the address keywords, determine the intersection coordinates of the two (i.e., the geographical location coordinates corresponding to the address feature information), and determine the POIs within a certain range of the intersection coordinates (this range can be determined as needed) as candidate geographical locations.

[0099] Sub-step S1044: Predict the target geographical location from at least one candidate geographical location, and determine the address word related to the address information according to the target geographical location.

[0100] Those skilled in the art can predict the target geographical location from the candidate geographical locations according to the needs by using appropriate rules.

[0101] For example, in sub-step S1044, predicting the target geographical location from at least one candidate geographical location can be implemented as: determining the target geographical location according to the query frequency of the candidate geographical locations. For example, select the candidate geographical location with the highest query frequency as the target geographical location.

[0102] In sub-step S1044, determining the address word related to the address information according to the target geographical location can be implemented as: generating an address word for indicating the target geographical location according to the target geographical location, as the address word related to the address information. For example, determine the name of the target geographical location as the address word related to the address information.

[0103] So as to generate an interaction statement for interacting with the user according to this address word later. If the user determines the address word in the interaction statement, the address structured information can be generated according to this address word for subsequent use.

[0104] The implementation process of the data processing method is described below in combination with the usage scenario of a specific dialogue system configured on the server side in a public security scenario:

[0105] Such as Figure 2b As shown, in this usage scenario, the dialogue system includes a semantic address parsing module, a geographical information database, and a standard address model.

[0106] Among them, the semantic address parsing module is used to extract the address feature information in the dialogue statement, predict the address word according to the POI returned by the geographical information database and generate the corresponding interaction statement to send to the user, and determine whether the predicted address word is accurate according to the user's dialogue statement. If it is accurate, send the address word to the standard address model to obtain the address structured information.

[0107] The geographic information database stores a vast amount of POI data (stored in the form of attributes, including but not limited to name, coordinates, query frequency, etc.). The geographic information database is used to search according to address feature information to obtain relevant POIs and return them to the semantic address parsing module. It can query specific geographic location coordinates and return all relevant POIs and their attributes at the queried geographic location coordinates, including the query frequency of the POIs, etc.

[0108] The standard address model is used to complete and structure the address words sent by the semantic address parsing module and return address structuring information.

[0109] Specifically, taking the example of a user reporting an emergency by phone in a public security scenario, the process of obtaining the user's address information (i.e., address structuring information) is as follows:

[0110] Process I: After answering the emergency call, play interaction statement 1, such as "Hello, may I ask where are you?", and obtain the user's dialogue statement 1, such as "The intersection of Pinghai Road and Zhonghe Middle Road". It should be noted that in this usage scenario, the interaction statement and the dialogue statement are taken as audio forms for illustration. In other usage scenarios, the interaction statement and the dialogue statement can be in text form, image form, audio-video form, etc., and this usage scenario is not restricted.

[0111] Process II: Determine whether dialogue statement 1 is a dialogue statement that confirms the address word in the previous interaction statement. If not, execute Processes III - VI; if so, execute Process VII. For example, input dialogue statement 1 into a trained classifier for semantic recognition to determine that it is not a dialogue statement that confirms the address word in the previous interaction statement, and then execute Processes III - VI; otherwise, execute Process VII.

[0112] Process III: Extract features from the address information in the dialogue statement to obtain address feature information. For example, the address feature information corresponding to dialogue statement 1 includes indication information indicating that the user's current location is at the intersection of Pinghai Road and Zhonghe Middle Road.

[0113] Process IV: Send the address feature information to the geographic information database, obtain the candidate geographic locations searched by the geographic information database, and determine the target geographic location based on the candidate geographic locations to predict the address word.

[0114] Send the address feature information (the indication information indicating that the user's current location is at the intersection of Pinghai Road and Zhonghe Middle Road) to the geographic information database. The geographic information database determines the coordinates of the intersection, searches for relevant POIs based on the coordinates of the intersection, and thus obtains the relevant POIs as candidate geographical locations (it should be noted that although the candidate geographical locations are used as examples for illustration, it is not limited to this, and other methods can be adopted as long as they can represent the corresponding POIs) and sends them to the semantic address parsing module.

[0115] The semantic address parsing module selects the POI corresponding to the candidate geographical location with the highest query frequency from the candidate geographical locations, and predicts the address word (such as the Metropole Hotel) based on the name of the POI.

[0116] Process V: Generate the interaction statement 2 (such as "Is it the Metropole Hotel?") according to the predicted address word and send it to the user.

[0117] For example, use a model that can automatically use natural language (such as the Seq2Seq model) to generate interaction statements. It consists of a sequence encoder and a sequence decoder. The sequence encoder is used to learn the user's input message, and the sequence decoder is used to generate a response message (i.e., the interaction statement) to the user based on the current state and chat history information. In this way, the interaction statement 2 can be automatically generated according to the address word.

[0118] Process VI: After obtaining the dialogue statement 2 in which the user replies to the interaction statement 2, return to Process II for execution. If the dialogue statement 2 is "opposite the Metropole Hotel", then execute Process II and determine it as no, then execute Processes III to VI, generate the interaction statement 3 (such as "Is it the Hangzhou Stomatological Hospital?") and obtain the dialogue statement 3 (such as "Yes"). Execute Process II again for the dialogue statement 3, determine it as yes, and execute Process VII.

[0119] Process VII: Perform address structuring on the address word in the previous interaction statement of the user's dialogue statement to obtain the address structuring information corresponding to the address word as the user's address information.

[0120] For example, the semantic address parsing module sends the predicted word (such as the Hangzhou Stomatological Hospital) in the previous interaction statement (i.e., the interaction statement 3) to the standard address model, and the standard address model performs completion processing and address structuring on it. Address structuring can be understood as performing structured parsing on the address segments in the address text, completing the annotation of place name elements and structured level mapping. Through address structuring, a series of address segments can be analyzed from the complete address information, that is, the address type annotation for each address segment (such as province, city, district, road, etc.) is generated, and the address structuring information (such as "Zhejiang Province (province), Hangzhou City (city), Shangcheng District (district), No. 1 Pinghai Road (street, building number), Hangzhou Stomatological Hospital (detailed address)") is sent to the semantic address parsing module.

[0121] The standard address model can be a neural network model trained using complete address data.

[0122] The semantic address parsing module can send the address structured information to the alarm center as the alarm target address information, so that the alarm center can obtain accurate and structured address information.

[0123] In this usage scenario, address information indicating the user's precise geographical location is obtained through an interactive dialogue. This solves the problem that in scenarios such as customer service and alarm, users usually initially return a simple and somewhat ambiguous short address that cannot be used. The semantic address parsing model automatically analyzes the ambiguous short address to generate a more accurate interactive statement containing address information, thereby guiding the user to provide more relevant address information. Through the repeated question-and-answer process until a definite, precise, and complete address information is obtained. This fully solves the problem in the existing technology that structured processing of the short address provided by the user cannot obtain accurate address information, resulting in the inability to determine the service entry and provide services.

[0124] Through this embodiment, feature extraction is performed on the user's dialogue statement, and address words are predicted based on the extracted address feature information to generate an interactive statement using the address words. Then, through the interactive statement, one or more rounds of dialogue are carried out with the user to gradually obtain more precise address information. This is more in line with human interaction habits and can obtain accurate address information even when the user's dialogue statement only includes an ambiguous short address, solving the problem in the existing technology that the user's location cannot be accurately determined using a short address.

[0125] In this way, it is realized that taking the dialogue of relevant address information as input, converting unstructured text into a structured address information representation, and through multiple rounds of dialogue with the user in combination with the query of the geographical information database, gradually obtaining more precise address information.

[0126] In addition, the Seq2Seq model can be used to train the part of the semantic address parsing module in the standard dialogue system for generating interactive statements, and CRF++ can be used for address structuring processing. When dealing with offline data involving large-scale data processing, a cloud computing platform for large-scale parallel computing can be used.

[0127] The data processing method of this embodiment can be executed by any suitable electronic device with data processing capabilities, including but not limited to: servers, mobile terminals (such as tablets, mobile phones, etc.) and PC machines, etc.

[0128] Embodiment III

[0129] Refer to Figure 3, showing a structural block diagram of a data processing device according to Embodiment 3 of the present invention.

[0130] The data processing device in this embodiment includes: a feature extraction module 302, configured to extract features from the address information in the user's dialogue statement to obtain the address feature information of the dialogue statement; a prediction module 304, configured to predict an address word related to the address information according to the address feature information; an interaction module 306, configured to generate an interaction statement for interacting with the user according to the address word, so as to determine the user's address information according to the dialogue statement in which the user responds to the interaction statement.

[0131] Optionally, the device further includes: a first determination module 300a, configured to determine whether the user's dialogue statement is a dialogue statement for confirming the address word in the previous interaction statement of the user's dialogue statement before the feature extraction module 302 extracts features from the address information in the user's dialogue statement; if not, the feature extraction module 302 performs the operation of extracting features from the address information in the user's dialogue statement.

[0132] Optionally, the device further includes: a processing module 300b, configured to perform address structuring processing on the address word in the previous interaction statement of the user's dialogue statement when the first determination module 300a determines it is, to obtain the address structuring information corresponding to the address word, and use the address structuring information as the user's address information.

[0133] Optionally, the processing module 300b is configured to perform address completion processing on the address word in the previous interaction statement of the user's dialogue statement to obtain complete address information; perform address structuring processing on the complete address information to obtain the address structuring information corresponding to the address word, and use the address structuring information as the user's address information.

[0134] Optionally, the address information in the user's dialogue statement includes an address keyword and an auxiliary word of the address keyword. The device further includes: an identification module 300c, configured to perform named entity recognition on the user's dialogue statement before the feature extraction module 302 extracts features from the address information in the user's dialogue statement, to obtain the address keyword and the auxiliary word of the address keyword in the user's dialogue statement.

[0135] Optionally, when the feature extraction module 302 extracts features from the address information in the user's dialogue statement, it matches an address feature extraction template for the address information according to the auxiliary word included in the address information; and extracts features from the address information according to the address feature extraction template to obtain the address feature information of the dialogue statement.

[0136] Optionally, the device further includes: a second determination module 308, configured to determine the commodity mailing address information according to the user's address information; or a third determination module 310, configured to determine the alarm target address information according to the user's address information.

[0137] Optionally, the prediction module 304 is configured to determine the geographical location coordinates corresponding to the address feature information according to the address feature information; search the preset geographical locations according to the geographical location coordinates corresponding to the address feature information to obtain at least one candidate geographical location; predict the target geographical location from the at least one candidate geographical location, and determine the address word related to the address information according to the target geographical location.

[0138] Optionally, before searching the preset geographical locations according to the geographical location coordinates corresponding to the address feature information, the prediction module 304 is further configured to determine the matching geographical location search strategy according to the address feature extraction template corresponding to the address feature information; when searching the preset geographical locations according to the geographical location coordinates corresponding to the address feature information, the prediction module 304 searches the preset geographical locations according to the geographical location coordinates corresponding to the address feature information through the geographical location search strategy to obtain at least one candidate geographical location.

[0139] Optionally, the address feature information includes the indication information for indicating that the user's geographical location is opposite to the reference geographical location indicated by the address keyword; when searching the preset geographical locations according to the geographical location coordinates corresponding to the address feature information through the geographical location search strategy, the prediction module 304 is configured to obtain the road direction information of the reference geographical location; search the preset geographical locations according to the direction information, the geographical location coordinates corresponding to the address feature information, and the indication information through the geographical location search strategy to obtain at least one candidate geographical location located opposite to the reference geographical location.

[0140] Optionally, the address feature information includes the indication information for indicating that the user's geographical location is at the intersection of two reference geographical locations indicated by the address keyword; when searching the preset geographical locations according to the geographical location coordinates corresponding to the address feature information through the geographical location search strategy, the prediction module 304 searches the preset geographical locations according to the geographical location coordinates corresponding to the address feature information and the indication information through the geographical location search strategy to obtain at least one candidate geographical location located at the intersection of the reference geographical locations.

[0141] Optionally, when predicting the target geographical location from the at least one candidate geographical location, the prediction module 304 is configured to determine the target geographical location according to the query frequency of the candidate geographical location.

[0142] Optionally, when determining an address term related to the address information according to the target geographical location, the prediction module 304 generates an address term for indicating the target geographical location according to the target geographical location, and uses it as the address term related to the address information.

[0143] The data processing device in this embodiment is used to implement the corresponding data processing methods in the foregoing multiple method embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be elaborated here. In addition, the function implementation of each module in the data processing device in this embodiment can refer to the description of the corresponding part in the foregoing method embodiments, which will not be elaborated here either.

[0144] Embodiment IV

[0145] Referring to Figure 4 , a schematic structural diagram of an electronic device according to Embodiment VI of the present invention is shown. The specific implementation of the electronic device in the specific embodiments of the present invention is not limited.

[0146] As Figure 4 shown, the electronic device may include: a processor 402, a communications interface 404, a memory 406, and a communication bus 408.

[0147] Wherein:

[0148] The processor 402, the communications interface 404, and the memory 406 communicate with each other through the communication bus 408.

[0149] The communications interface 404 is used to communicate with other electronic devices such as terminal devices or servers.

[0150] The processor 402 is used to execute the program 410, and specifically can execute the relevant steps in the foregoing data processing method embodiments.

[0151] Specifically, the program 410 may include program code, and the program code includes computer operation instructions.

[0152] The processor 402 may be a central processing unit CPU, or a specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present invention. One or more processors included in the electronic device may be of the same type of processor, such as one or more CPUs; or may be of different types of processors, such as one or more CPUs and one or more ASICs.

[0153] A memory 406 for storing a program 410. The memory 406 may include high-speed RAM memory and may also include non-volatile memory, such as at least one disk memory.

[0154] The program 410 may specifically be used to cause the processor 402 to perform the following operations: extract features from the address information in the user's dialogue statement to obtain the address feature information of the dialogue statement; predict an address word related to the address information according to the address feature information; generate an interaction statement for interacting with the user according to the address word, so as to determine the user's address information according to the dialogue statement in response to the interaction statement by the user.

[0155] In an alternative embodiment, the program 410 is further used to cause the processor 402 to determine whether the user's dialogue statement is a dialogue statement for confirming the address word in the previous interaction statement of the user's dialogue statement before extracting features from the address information in the user's dialogue statement; if not, then perform the operation of extracting features from the address information in the user's dialogue statement.

[0156] In an alternative embodiment, the program 410 is further used to cause the processor 402, if so, to perform address structuring processing on the address word in the previous interaction statement of the user's dialogue statement to obtain the address structuring information corresponding to the address word, and use the address structuring information as the user's address information.

[0157] In an alternative embodiment, the program 410 is further used to cause the processor 402, when performing address structuring processing on the address word in the previous interaction statement of the user's dialogue statement, perform address completion processing on the address word in the previous interaction statement of the user's dialogue statement to obtain complete address information; perform address structuring processing on the complete address information to obtain the address structuring information corresponding to the address word, and use the address structuring information as the user's address information.

[0158] In an alternative embodiment, the address information in the user's dialogue statement includes an address keyword and an auxiliary word of the address keyword. The program 410 is further used to cause the processor 402 to perform named entity recognition on the user's dialogue statement before extracting features from the address information in the user's dialogue statement, so as to obtain the address keyword and the auxiliary word of the address keyword in the user's dialogue statement.

[0159] In an alternative embodiment, the program 410 is further used to cause the processor 402, when extracting features from the address information in the user's dialogue statement, match an address feature extraction template for the address information according to the auxiliary word included in the address information; extract features from the address information according to the address feature extraction template to obtain the address feature information of the dialogue statement.

[0160] In an alternative embodiment, the program 410 is further configured to cause the processor 402 to determine the commodity mailing address information according to the user's address information; or determine the alarm target address information according to the user's address information.

[0161] In an alternative embodiment, the program 410 is further configured to cause the processor 402, when predicting an address word related to the address information according to the address feature information, to determine the geographical location coordinates corresponding to the address feature information according to the address feature information; search for a preset geographical location according to the geographical location coordinates corresponding to the address feature information to obtain at least one candidate geographical location; predict a target geographical location from the at least one candidate geographical location, and determine an address word related to the address information according to the target geographical location.

[0162] In an alternative embodiment, the program 410 is further configured to cause the processor 402 to determine a matching geographical location search strategy according to the address feature extraction template corresponding to the address feature information before searching for a preset geographical location according to the geographical location coordinates corresponding to the address feature information; the program 410 is further configured to cause the processor 402 to search for a preset geographical location according to the geographical location coordinates corresponding to the address feature information through the geographical location search strategy when searching for a preset geographical location according to the geographical location coordinates corresponding to the address feature information to obtain at least one candidate geographical location.

[0163] In an alternative embodiment, the address feature information includes indication information for indicating that the geographical location of the user is opposite to the reference geographical location indicated by the address keyword; the program 410 is further configured to cause the processor 402 to obtain the road direction information of the reference geographical location when searching for a preset geographical location according to the geographical location coordinates corresponding to the address feature information through the geographical location search strategy; search for a preset geographical location according to the direction information, the geographical location coordinates corresponding to the address feature information, and the indication information through the geographical location search strategy to obtain at least one candidate geographical location located opposite to the reference geographical location.

[0164] In an alternative embodiment, the address feature information includes indication information for indicating that the geographical location of the user is at the intersection of two reference geographical locations indicated by the address keyword; the program 410 is further configured to cause the processor 402 to search for a preset geographical location according to the geographical location coordinates corresponding to the address feature information and the indication information through the geographical location search strategy when searching for a preset geographical location according to the geographical location coordinates corresponding to the address feature information through the geographical location search strategy to obtain at least one candidate geographical location located at the intersection of the reference geographical locations.

[0165] In an alternative embodiment, the program 410 is further configured to cause the processor 402 to determine a target geographical location according to the query frequency of the candidate geographical locations when predicting the target geographical location from at least one candidate geographical location.

[0166] In an alternative embodiment, the program 410 is further configured to cause the processor 402 to generate an address term for indicating the target geographical location as the address term related to the address information according to the target geographical location when determining the address term related to the address information according to the target geographical location.

[0167] For the specific implementation of each step in the program 410, reference may be made to the corresponding steps and descriptions in the corresponding units in the foregoing data processing method embodiments, which will not be elaborated herein. Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the devices and modules described above can refer to the corresponding process descriptions in the foregoing method embodiments, which will not be repeated herein.

[0168] It should be noted that, according to the needs of implementation, each component / step described in the embodiments of the present invention can be split into more components / steps, or two or more components / steps or partial operations of components / steps can be combined into new components / steps to achieve the purpose of the embodiments of the present invention.

[0169] The method according to the embodiments of the present invention can be implemented in hardware, firmware, or be implemented as software or computer code that can be stored in a recording medium (such as a CD ROM, RAM, floppy disk, hard disk, or magneto-optical disk), or be implemented as computer code originally stored in a remote recording medium or a non-transitory machine-readable medium and to be downloaded through a network and stored in a local recording medium, so that the method described herein can be stored in such a software process on a recording medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware (such as an ASIC or FPGA). It can be understood that a computer, a processor, a microprocessor controller, or programmable hardware includes a storage component (such as RAM, ROM, flash memory, etc.) that can store or receive software or computer code, and when the software or computer code is accessed and executed by the computer, the processor, or the hardware, the data processing method described herein is implemented. In addition, when a general-purpose computer accesses the code for implementing the data processing method shown herein, the execution of the code converts the general-purpose computer into a dedicated computer for executing the data processing method shown herein.

[0170] Those of ordinary skill in the art will appreciate that the units and method steps of each example described in connection with the embodiments disclosed herein can be implemented using electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Skilled artisans can use different methods for each specific application to implement the described functions, but such implementation should not be considered to exceed the scope of the embodiments of the present invention.

[0171] The above embodiments are only used to illustrate the embodiments of the present invention, rather than to limit the embodiments of the present invention. Those of ordinary skill in the relevant technical field can also make various changes and modifications without departing from the spirit and scope of the embodiments of the present invention. Therefore, all equivalent technical solutions also belong to the scope of the embodiments of the present invention. The scope of patent protection of the embodiments of the present invention shall be defined by the claims.

Claims

1. A data processing method, comprising: extracting features of address information in a user's dialogue statement to obtain address feature information of the dialogue statement; predicting an address word related to the address information according to the address feature information; generating an interaction statement for interacting with the user according to the address word, so as to determine the user's address information according to the dialogue statement in response to the interaction statement by the user; wherein, the address information in the dialogue statement includes an address keyword and an auxiliary word of the address keyword, and the address feature information includes indication information for representing the corresponding relationship between the position where the user is currently located and the reference geographical location characterized by the address keyword; wherein, the predicting an address word related to the address information according to the address feature information includes: determining geographical position coordinates corresponding to the address feature information according to the address feature information; searching for a preset geographical position according to the geographical position coordinates corresponding to the address feature information to obtain at least one candidate geographical position, wherein the candidate geographical position is used to represent an interest point that meets the requirements of the geographical position coordinates corresponding to the address feature information; predicting a target geographical position from the at least one candidate geographical position, and determining an address word related to the address information according to the target geographical position.

2. The method according to claim 1, wherein Before extracting features of address information in the user's dialogue statement, it further includes: determining whether the user's dialogue statement is a dialogue statement for confirming an address word in a previous interaction statement of the user's dialogue statement; if not, then performing the operation of extracting features of address information in the user's dialogue statement.

3. The method according to claim 2, wherein The method further includes: if so, performing address structuring processing on the address word in the previous interaction statement of the user's dialogue statement to obtain address structuring information corresponding to the address word, and using the address structuring information as the user's address information.

4. The method according to claim 3, wherein, The performing address structuring processing on the address word in the previous interaction statement of the user's dialogue statement includes: performing address completion processing on the address word in the previous interaction statement of the user's dialogue statement to obtain complete address information; performing address structuring processing on the complete address information to obtain address structuring information corresponding to the address word, and using the address structuring information as the user's address information.

5. The method according to claim 1, wherein Before extracting features of address information in the user's dialogue statement, it further includes: performing named entity recognition on the user's dialogue statement to obtain the address keyword and the auxiliary word of the address keyword in the user's dialogue statement.

6. The method according to claim 5, wherein, The extracting features of address information in the user's dialogue statement includes: matching an address feature extraction template for the address information according to the auxiliary word included in the address information; extracting features of the address information according to the address feature extraction template to obtain address feature information of the dialogue statement.

7. The method according to claim 1, wherein The method further includes: determining commodity mailing address information according to the user's address information; or, Determine the alarm target address information according to the address information of the user.

8. The method according to claim 1, wherein Before searching the preset geographical locations according to the geographical location coordinates corresponding to the address feature information, it further includes: Determine the matching geographical location search strategy according to the address feature extraction template corresponding to the address feature information; The searching the preset geographical locations according to the geographical location coordinates corresponding to the address feature information includes: Through the geographical location search strategy, search the preset geographical locations according to the geographical location coordinates corresponding to the address feature information to obtain at least one candidate geographical location.

9. The method according to claim 8, wherein The address feature information includes indication information for indicating that the geographical location of the user is opposite to the reference geographical location indicated by the address keyword; The searching the preset geographical locations through the geographical location search strategy according to the geographical location coordinates corresponding to the address feature information includes: Obtain the direction information of the road where the reference geographical location is located; Through the geographical location search strategy, search the preset geographical locations according to the direction information, the geographical location coordinates corresponding to the address feature information, and the indication information to obtain at least one candidate geographical location opposite to the reference geographical location.

10. The method according to claim 8, wherein The address feature information includes indication information for indicating that the geographical location of the user is at the intersection of two reference geographical locations indicated by the address keyword; The searching the preset geographical locations through the geographical location search strategy according to the geographical location coordinates corresponding to the address feature information includes: Through the geographical location search strategy, search the preset geographical locations according to the geographical location coordinates corresponding to the address feature information and the indication information to obtain at least one candidate geographical location at the intersection of the reference geographical locations.

11. The method according to claim 1, wherein The predicting the target geographical location from the at least one candidate geographical location includes: Determine the target geographical location according to the query frequency of the candidate geographical location.

12. The method according to claim 1, characterized in that, The determining the address word related to the address information according to the target geographical location includes: Generate an address word for indicating the target geographical location according to the target geographical location as the address word related to the address information.

13. A data processing device, comprising: A feature extraction module, configured to extract features from the address information in the user's dialogue statement to obtain the address feature information of the dialogue statement; A prediction module, configured to predict an address word related to the address information according to the address feature information; An interaction module, configured to generate an interaction statement for interacting with the user according to the address word, so as to determine the address information of the user according to the dialogue statement of the user in response to the interaction statement; Wherein, the address information in the dialogue statement includes an address keyword and an auxiliary word of the address keyword, and the address feature information includes indication information for representing the corresponding relationship between the current location of the user and the reference geographical location characterized by the address keyword; Among them, the prediction module is used to predict an address word related to the address information according to the address feature information through the following steps: determining a geographical location coordinate corresponding to the address feature information according to the address feature information; searching for a preset geographical location according to the geographical location coordinate corresponding to the address feature information to obtain at least one candidate geographical location, where the candidate geographical location is used to represent an interest point that meets the requirements of the geographical location coordinate corresponding to the address feature information; predicting a target geographical location from the at least one candidate geographical location, and determining an address word related to the address information according to the target geographical location.

14. An electronic device, comprising: A processor, a memory, a communication interface, and a communication bus, where the processor, the memory, and the communication interface complete communication with each other through the communication bus; The memory is used to store at least one executable instruction, and the executable instruction causes the processor to execute an operation corresponding to the data processing method according to any one of claims 1-12.

15. A computer storage medium, on which a computer program is stored, and when the program is executed by a processor, it implements the data processing method according to any one of claims 1-12.

Images