CN112860878A

CN112860878A - Service data recommendation method, storage medium and equipment

Info

Publication number: CN112860878A
Application number: CN202110160967.6A
Authority: CN
Inventors: 谢家欢; 张懿; 刘冬冬; 陈文涛; 杨泽洋
Original assignee: Weimin Insurance Agency Co Ltd
Current assignee: Weimin Insurance Agency Co Ltd
Priority date: 2021-02-05
Filing date: 2021-02-05
Publication date: 2021-05-28

Abstract

The embodiment of the application discloses a service data recommendation method, a storage medium and equipment. The method comprises the following steps: the first equipment responds to the information input operation aiming at the target service and sends the service consultation information determined by the information input operation to the second equipment; the second equipment acquires M target material data matched with the service consultation information in the target service and displays the M target material data, wherein M is a positive integer, the second equipment responds to selection operation aiming at the M target material data, the target material data determined by the selection operation is determined as consultation feedback information in the M target material data, and the second equipment sends the consultation feedback information to the first equipment. By the method and the device, the accuracy and efficiency of service data recommendation can be improved.

Description

Service data recommendation method, storage medium and equipment

Technical Field

The present application relates to the field of data processing technologies, and in particular, to a service data recommendation method, a storage medium, and a device.

Background

With the rapid development of various industries in recent years, the number of service businesses provided by various industries for users is increasing day by day, the number is huge, how to rapidly and accurately provide consultation services for users and help users to develop demand businesses becomes a problem to be solved by organizations/personnel of various industries.

In the prior art, when a user consults a question on line, relevant service personnel are required to manually inquire so as to obtain an answer to the question of the user and reply the answer to the user; however, the related service personnel usually need to face a lot of advisory work, and it takes a lot of time to manually query the answers to the questions, thereby resulting in too low processing efficiency of the advisory service; moreover, the answers inquired by the service personnel have personal subjectivity, and may have a certain difference with the answer wanted by the user, so that the reply accuracy of the consultation service is too low.

Disclosure of Invention

The technical problem to be solved by the embodiments of the present application is to provide a service data recommendation method, a storage medium, and a device, which can improve accuracy and efficiency of service data recommendation.

An embodiment of the present application provides a method for recommending service data, including:

the first equipment responds to the information input operation aiming at the target service and sends the service consultation information determined by the information input operation to the second equipment;

the second equipment acquires M target material data matched with the service consultation information in the target service and displays the M target material data; m is a positive integer;

the second equipment responds to selection operation aiming at the M target material data, and determines the target material data determined by the selection operation as consultation feedback information in the M target material data;

the second device transmits the advisory feedback information to the first device.

The second device obtains M pieces of target material data matched with the service consultation information in the target service, and the method comprises the following steps:

the second equipment receives the business consultation information sent by the first equipment and acquires an information characteristic vector corresponding to the business consultation information;

the second equipment acquires intention probabilities between the information feature vectors and the N candidate intentions respectively; n is a positive integer;

the second equipment determines a target intention corresponding to the business consultation information from the N candidate intentions according to the intention probability;

and the second equipment acquires M target material data matched with the target intention in the target service.

The obtaining of the information feature vector corresponding to the service consultation information includes:

the second equipment acquires the consultation keywords in the service consultation information, and performs vector conversion on the consultation keywords to obtain word vectors corresponding to the consultation keywords;

and the second equipment acquires the information characteristic vector corresponding to the consultation keyword according to the semantic information corresponding to the word vector.

The second device obtains intention probabilities between the information feature vectors and the N candidate intentions respectively, and the intention probabilities include:

the second equipment inputs the information characteristic vector to a neural network layer in the intention recognition model, and obtains a prediction vector corresponding to the information characteristic vector according to a weight matrix corresponding to the neural network layer;

the second device determines intent probabilities between the prediction vectors and the N candidate intentions, respectively, in a classifier of the intent recognition model.

Wherein, show M target material data, include:

the second equipment acquires material attribute characteristics corresponding to the M pieces of target material data respectively, and acquires user attribute characteristics corresponding to target users providing business consultation information;

the second equipment determines recommended evaluation values corresponding to the M target material data respectively according to the material attribute characteristics, the user attribute characteristics and the information characteristic vectors corresponding to the service consultation information;

and the second equipment sorts the M recommended evaluation values and displays the M target material data according to the sorted M recommended evaluation values.

The material attribute characteristics comprise material semantic characteristics and material type characteristics;

the second device determines recommended evaluation values corresponding to the M target material data respectively according to the material attribute characteristics, the user attribute characteristics and the information characteristic vector corresponding to the service consultation information, and the recommended evaluation values include:

the second equipment inputs the material semantic features, the user attribute features and the information feature vectors corresponding to the service consultation information into a recommendation model, and determines prediction evaluation values corresponding to M target material data respectively according to the recommendation model;

the second equipment acquires weight coefficients corresponding to the M target material data respectively according to the material type characteristics;

and the second equipment determines recommended evaluation values corresponding to the M target material data respectively according to the weight coefficient and the predicted evaluation value.

the first equipment responds to information input operation aiming at the target service, and sends the service consultation information determined by the information input operation to the second equipment so that the second equipment determines consultation feedback information corresponding to the service consultation information in M target material data contained in the target service; m is a positive integer;

and the first equipment receives the consultation feedback information sent by the second equipment and displays the consultation feedback information.

Wherein, the method also comprises:

the first equipment acquires user behavior data aiming at the consultation feedback information and sends the user behavior data to the second equipment so that the second equipment updates the recommendation model based on the user behavior data; the recommendation model is used for determining the prediction evaluation values corresponding to the M target material data, and the prediction evaluation values provide basis conditions for the second equipment to determine the consultation feedback information.

An aspect of an embodiment of the present application provides a service data recommendation device, including:

the first sending module is used for responding to the information input operation aiming at the target service and sending the service consultation information determined by the information input operation to the second equipment;

the display module is used for acquiring M target material data matched with the service consultation information in the target service and displaying the M target material data; m is a positive integer;

the first determining module is used for responding to the selection operation aiming at the M target material data, and determining the target material data determined by the selection operation as the consultation feedback information in the M target material data;

and the second sending module is used for sending the consultation feedback information to the first equipment.

Wherein the first determining module comprises:

the first acquisition unit is used for receiving the service consultation information sent by the first equipment and acquiring an information characteristic vector corresponding to the service consultation information;

a second obtaining unit, configured to obtain intention probabilities between the information feature vectors and the N candidate intentions, respectively; n is a positive integer;

the first determining unit is used for determining a target intention corresponding to the business consultation information from the N candidate intentions according to the intention probability;

and the third acquisition unit is used for acquiring M target material data matched with the target intention in the target service.

The first obtaining unit is specifically configured to:

acquiring consultation keywords in the service consultation information, and performing vector conversion on the consultation keywords to obtain word vectors corresponding to the consultation keywords;

and obtaining the information characteristic vector corresponding to the consultation keyword according to the semantic information corresponding to the word vector.

The second obtaining unit is specifically configured to:

inputting the information characteristic vector into a neural network layer in the intention recognition model, and acquiring a prediction vector corresponding to the information characteristic vector according to a weight matrix corresponding to the neural network layer;

in a classifier of an intent recognition model, intent probabilities between prediction vectors and N candidate intentions, respectively, are determined.

Wherein, the display module includes:

a fourth obtaining unit, configured to obtain material attribute features corresponding to the M pieces of target material data, and obtain a user attribute feature corresponding to a target user providing service consultation information;

the second determining unit is used for determining recommended evaluation values corresponding to the M target material data respectively according to the material attribute characteristics, the user attribute characteristics and the information characteristic vectors corresponding to the service consultation information;

and the display unit is used for sequencing the M recommended evaluation values and displaying the M target material data according to the sequenced M recommended evaluation values.

the second determining unit is specifically configured to:

inputting the material semantic features, the user attribute features and the information feature vectors corresponding to the service consultation information into a recommendation model, and determining prediction evaluation values corresponding to M target material data respectively according to the recommendation model;

acquiring weight coefficients corresponding to M pieces of target material data respectively according to the material type characteristics;

and determining recommended evaluation values corresponding to the M target material data respectively according to the weight coefficient and the predicted evaluation value.

the second determining module is used for responding to information input operation aiming at the target service, and sending the service consultation information determined by the information input operation to the second equipment so that the second equipment determines consultation feedback information corresponding to the service consultation information in M target material data contained in the target service; m is a positive integer;

and the receiving module is used for receiving the transmitted consultation feedback information and displaying the consultation feedback information.

Wherein, the device still includes:

the third sending module is used for acquiring user behavior data aiming at the consultation feedback information and sending the user behavior data to the second equipment so that the second equipment can update the recommendation model based on the user behavior data; the recommendation model is used for determining the prediction evaluation values corresponding to the M target material data, and the prediction evaluation values provide basis conditions for the second equipment to determine the consultation feedback information.

One aspect of the present application provides a computer device, comprising: a processor and a memory;

wherein, the memorizer is used for storing the computer program, the processor is used for calling the above-mentioned computer program, in order to carry out the following step:

responding to the information input operation aiming at the target service, and sending the service consultation information determined by the information input operation to the second equipment;

acquiring M target material data matched with the service consultation information in the target service, and displaying the M target material data; m is a positive integer;

responding to selection operation aiming at the M target material data, and determining the target material data determined by the selection operation as consultation feedback information in the M target material data;

and sending the consultation feedback information to the first equipment.

An aspect of the embodiments of the present application provides a computer-readable storage medium, where a computer program is stored, and when the computer program is executed by a processor, the computer program performs the following steps:

and sending the consultation feedback information to the first equipment.

An aspect of the application provides a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions to cause the computer device to perform the method of the above-described aspect.

In the embodiment of the application, the first device responds to the information input operation aiming at the target service, and the service consultation information determined by the information input operation is sent to the second device. And the second equipment acquires M target material data matched with the service consultation information in the target service and displays the M target material data, wherein M is a positive integer. The second equipment responds to selection operation aiming at the M target material data, determines the target material data determined by the selection operation as consultation feedback information in the M target material data, and sends the consultation feedback information to the first equipment. It can be seen that, for the service consultation information sent by the first device, the second device can preliminarily screen M target material data matched with the service consultation information in the target service, and display the M target material data, consultation feedback information corresponding to the service consultation information can be determined from the displayed M target material data, and the consultation feedback information is returned to the first device as reply information of the service consultation information, so that tedious manual operation can be reduced, the recommendation efficiency of the service data can be improved, and the accuracy of service data recommendation can be improved by using the consultation feedback information determined from the preliminarily screened M target material data.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.

Fig. 1 is a schematic architecture diagram of a business data recommendation system provided in an embodiment of the present application;

fig. 2 is a schematic flowchart of a service data recommendation method according to an embodiment of the present application;

FIG. 3 is a schematic diagram of an intent recognition model provided by an embodiment of the present application;

FIG. 4 is a block diagram of a sensing layer in a neural network layer provided in an embodiment of the present application;

fig. 5 is a schematic diagram illustrating counseling feedback information corresponding to service counseling information according to an embodiment of the present disclosure;

FIG. 6 is a schematic diagram illustrating a method for recommending advisory feedback information for insurance services for a target user according to an embodiment of the present application;

fig. 7 is a schematic diagram illustrating a method for determining advisory feedback information in the related art according to an embodiment of the present application;

fig. 8a is a schematic diagram of target material data determined to be displayed according to an embodiment of the present application;

fig. 8b is a schematic diagram of determining advisory feedback information according to an embodiment of the present application;

fig. 8c is a schematic diagram of determining advisory feedback information according to an embodiment of the present application;

fig. 8d is a schematic diagram of target material data provided by an embodiment of the present application;

fig. 9 is a schematic diagram of a service data recommendation method according to an embodiment of the present application;

fig. 10 is a schematic diagram for displaying advisory feedback information provided by an embodiment of the present application;

fig. 11 is a schematic structural diagram of a service data recommendation device according to an embodiment of the present application

Fig. 12 is a schematic structural diagram of a service data recommendation device according to an embodiment of the present application;

fig. 13 is a schematic structural diagram of a computer device according to an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are only a part of the embodiments of the present application, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

Referring to fig. 1, fig. 1 is a schematic structural diagram of a business data recommendation system provided in an embodiment of the present application. As shown in fig. 1, the business data recommendation system may include a server 10 and a user terminal cluster. The user terminal cluster may comprise one or more user terminals, where the number of user terminals will not be limited. As shown in fig. 1, the system may specifically include a user terminal 100a, a user terminal 100b, user terminals 100c and …, and a user terminal 100 n. As shown in fig. 1, the user terminal 100a, the user terminal 100b, the user terminals 100c, …, and the user terminal 100n may be respectively connected to the server 10 via a network, so that each user terminal may interact with the server 10 via the network.

Wherein, each ue in the ue cluster may include: the intelligent terminal comprises an intelligent terminal with service data recommendation, such as a smart phone, a tablet computer, a notebook computer, a desktop computer, wearable equipment, an intelligent home, and head-mounted equipment. It should be understood that each user terminal in the user terminal cluster shown in fig. 1 may be installed with a target application (i.e., an application client), and when the application client runs in each user terminal, data interaction may be performed with the server 10 shown in fig. 1.

As shown in fig. 1, the server 10 may transmit the service consultation information and the consultation feedback information, and acquire M pieces of target material data matching the service consultation information in the target service. The server 10 may be an independent physical server, a server cluster or a distributed system formed by a plurality of physical servers, or a cloud server providing basic cloud computing services such as a cloud service, a cloud database, cloud computing, a cloud function, cloud storage, a network service, cloud communication, a middleware service, a domain name service, a security service, a CDN, a big data and artificial intelligence platform, and the like.

In this embodiment, the first device and the second device may refer to any two of the user terminal 100a, the user terminal 100b, the user terminals 100c, …, and the user terminal 100n, and for convenience of understanding, the first device in this embodiment may refer to the user terminal 100a shown in fig. 1, and the second device may refer to the user terminal 100b shown in fig. 1. The user terminal 100a and the user terminal 100b may both integrate a target application (i.e., an application client) with a service data recommendation function, the server 10 at this time may be understood as a background server of the target application, and data interaction may be performed between the user terminal 100a and the user terminal 100b through the background server corresponding to the target application. For example, the user a may input the service query information in the target application installed in the user terminal 100a, the user a at this time may be referred to as a target user, the user terminal 100a may obtain the service query information input by the user a and send the service query information to the server 10, after receiving the service query information sent by the user terminal 100a, the server 10 may identify M target material data matched with the service query information in the target service corresponding to the service query information and send the M target material data to the user terminal 100b, where M is a positive integer, and if M may be 1, 2, 3, … …. After receiving the M pieces of target material data transmitted from the server 10, the user terminal 100b can display the M pieces of target material in the target application. Of course, the server 10 may transmit the service counseling information to the user terminal 100b after receiving the service counseling information, and thus may display the service counseling information in the target application of the user terminal 100 b. It is understood that, after receiving the service counseling information, the server 10 may identify M pieces of target material data matching the service counseling information while transmitting the service counseling information to the user terminal 100b, and thus, the time of receiving the service counseling information may be earlier than the time of receiving the M pieces of target material data for the user terminal 100 b.

Further, for the M pieces of target material data displayed in the target application of the user terminal 100b, the holding user (also referred to as a service person) of the user terminal 100b may select data that is most matched with the service consultation information from the M pieces of target material data as consultation feedback information, the user terminal 100b may respond to the selection operation of the service person to obtain consultation feedback information corresponding to the service consultation information, transmit the consultation feedback information to the user terminal 100a through the server 10, and the user terminal 100a displays the consultation feedback information to provide consultation answering service for the user a.

Optionally, the user a and the service staff may communicate through different applications, that is, the target application installed on the user terminal 100a may be a first application, the target application installed on the user terminal 100b may be a second application, and the server 10 at this time may include a background server corresponding to the first application and a background server corresponding to the second application (the user terminal 100a where the first application is installed and the background server of the first application may be referred to as a first device, and the user terminal 100b where the second application is installed and the background server of the second application may be referred to as a second device). For example, the user a may input the service consultation information in the first application installed in the user terminal 100a, the user terminal 100a may obtain the service consultation information input by the user a, and send the service consultation information to the background server of the first application, the background server of the first application may further transmit the service consultation information to the background server of the second application, the background server of the second application may transmit the service consultation information to the user terminal 100b upon receiving the service consultation information, and display the service consultation information in the second application of the user terminal 100 b; meanwhile, the background server of the second application may identify M pieces of target material data matched with the service counseling information in the target service corresponding to the counseling service information, and the background server of the second application may transmit the M pieces of target material data to the user terminal 100 b. The user terminal 100b may display M pieces of target material data in the second application, and the M pieces of target material data may be displayed in different areas of the same interface with the service counseling information. The service staff can select data which is most matched with the service consultation information from the M target material data as consultation feedback information, the user terminal 100b can respond to the selection operation of the service staff to obtain the consultation feedback information corresponding to the service consultation information, the consultation feedback information is transmitted to the background server of the second application, and then the consultation feedback information can be transmitted to the user terminal 100a through data transmission between the background server of the second application and the background server of the first application, the background server of the first application transmits the consultation feedback information to the user terminal 100a, and the user terminal 100a displays the consultation feedback information to provide consultation answering service for the user A.

Referring to fig. 2, fig. 2 is a schematic flowchart of a service data recommendation method according to an embodiment of the present application. The service data recommendation method may be interactively executed by a first device and a second device, where the first device and the second device may be a server (such as the server 10 in fig. 1), or a user terminal (such as any user terminal in the user terminal cluster in fig. 1), or a system formed by the server and the user terminal, and this application is not limited thereto. As shown in fig. 2, the service data recommendation method may include steps S101-S104.

S101, the first equipment responds to the information input operation aiming at the target service and sends the service consultation information determined by the information input operation to the second equipment.

Specifically, in a service consultation scenario, when a user is confused about a service in a target application, service consultation information for the service may be input on an interactive interface corresponding to the used first device. The first device may respond to an information input operation of the target user for the target service, obtain service consultation information input by the target user for the target service, and send the service consultation information determined by the information input operation to the second device. The target application may refer to an application client for providing service data recommendation service, such as an application for providing service consultation service, e.g., an electronic mall application, a communication application, an insurance application, and the like; the target service may refer to a shopping service in an electronic mall, a package service in a communication application, an insurance service in an insurance application, and the like, and the embodiments of the present application do not limit the types of the target application and the target service.

S102, the second equipment acquires M target material data matched with the service consultation information in the target service and displays the M target material data; m is a positive integer.

Specifically, after receiving the service consultation information for the target service sent by the first device, the second device may obtain M target material data matched with the service consultation information in the target service, and display the M target material data on an interactive interface corresponding to the second device.

Optionally, the specific manner of acquiring, by the second device, the M pieces of target material data matched with the service consultation information in the target service may include: and the second equipment receives the service consultation information sent by the first equipment and acquires the information characteristic vector corresponding to the service consultation information. The second equipment obtains intention probabilities between information feature vectors corresponding to the business consultation information and the N candidate intentions respectively, wherein N is a positive integer, the second equipment determines a target intention corresponding to the business consultation information from the N candidate intentions according to the intention probabilities, and the second equipment obtains M target material data matched with the target intention in the target business.

Specifically, after receiving the service consultation information sent by the first device, the second device may obtain an information feature vector corresponding to the service consultation information, for example, the feature vector conversion may be performed on the service consultation information through a feature vector conversion model, so as to obtain an information feature vector composed of floating point numbers and corresponding to the service consultation information. After the second device obtains the information feature vector corresponding to the service consultation information, the intention probability between the information feature vector and the N candidate intentions can be obtained, namely, the intention information corresponding to the service consultation information is judged, and the target intention of the target user for the target service is determined. After the second device obtains the intention probabilities between the business consultation information and the N candidate intentions, the second device may determine a target intention corresponding to the business consultation information from the N candidate intentions according to the intention probability between the business consultation information and each candidate intention. Since there are various meanings between each person's speaking style and words, business consultation information for a target business inputted by a target user may belong to intention a or intention B. Therefore, intention probabilities between the business consultation information and the N candidate intentions respectively can be obtained, and the target intention corresponding to the business consultation information is determined from the N candidate intentions according to the intention probabilities. And if the intention probability corresponding to each candidate intention is ranked, ranking the candidate intents with the intention probability being the first three, and determining the candidate intents as the target intents corresponding to the business consultation information. After the second device obtains the target intention, M target material data matched with the target intention can be obtained in the target service, and the M material data are used for solving the service consultation information of the target user for the target service.

Optionally, the specific manner of obtaining the information feature vector corresponding to the service consultation information by the second device may include: the second equipment obtains the consultation keywords in the business consultation information, performs vector conversion on the consultation keywords to obtain word vectors corresponding to the consultation keywords, and obtains information characteristic vectors corresponding to the consultation keywords according to the semantic information corresponding to the word vectors.

Specifically, the second device may obtain a consultation keyword in the service consultation information, where the consultation keyword may be a keyword, or a keyword sentence, and the like in the service consultation information, that is, the keyword in the service consultation information of the target user for the target service is extracted, and the service consultation information of the target user for the target service is analyzed. As in insurance business, business consultation information of target users for target business may refer to "those contents of accident insurance support", "those can be purchased for a serious insurance in one year", and so on. When the keywords in the consulting service information are obtained, the candidate keywords in the consulting service information can be extracted, and then the candidate keywords are replaced by the standard keywords pre-stored in the second equipment, so that the consulting keywords corresponding to the consulting service information are obtained. The business consultation information for insurance business input by the target user is' buying heavy insurance, how do i need to operate? "the second device gets the business consultation information" buy heavy insurance, how do i need to operate? After that, candidate keywords such as "purchase", "heavy insurance", "operation", etc. in the business consultation information can be extracted. And acquiring standard keywords matched with the candidate keywords from the keyword database according to the candidate keywords corresponding to the service consultation information to serve as consultation keywords. For example, "buy a critical illness, how do i need to do? "how do the corresponding consulting keywords be" how to buy the heavy insurance? ".

After the second device obtains the consulting keywords corresponding to the service consulting information, vector conversion can be performed on the consulting keywords to obtain word vectors corresponding to the consulting keywords, and the information feature vectors corresponding to the consulting keywords are obtained according to the semantic information corresponding to the word vectors. The method comprises the steps of obtaining a word vector corresponding to the consultation keyword, and converting the consultation keyword into a vector through a vector conversion model. Specifically, the vector conversion model may be a BERT model (language representation model), where the BERT model is to obtain a vector in which full-text semantic information is fused with each word/word in an input text. After the consultation keywords in the service consultation information are obtained, the consultation keywords are input into the BERT model, and the BERT model can convert the consultation keywords into one-dimensional word vectors by inquiring the word vector table. The word vector table is a word vector corresponding to the feature information of each word collected in advance. Meanwhile, when a plurality of consultation keywords are available, the BERT model can also obtain the corresponding global semantic vectors among the consultation keywords and is fused with the word vectors corresponding to the consultation keywords. In addition, because semantic information carried by words/phrases appearing at different positions is different, the BERT model can also obtain position vectors corresponding to the consultation keywords, that is, the BERT model can obtain position vectors corresponding to the consultation keywords at different positions. Like "I love you" and "you love me", the same is formed by "I", "love", "you" three words, but because the positional information of the word is different, the semantic information that characterizes is different. And the BERT model obtains the sentence characteristic vector corresponding to the consultation keyword according to the word vector, the global semantic vector and the position vector corresponding to the consultation keyword.

The BERT model in the embodiment of the present application is a vector transformation model obtained by fine-tuning a basic BERT model, that is, a current corpus corresponding to a target service (that is, a service corpus corresponding to the target service) is added to the basic model for fine-tuning. Specifically, the specific manner of obtaining the BERT model in the embodiment of the present application may include: the method includes the steps of obtaining an initial vector conversion model (namely, a basic BERT model), training sample data and a label vector corresponding to the training sample data, wherein label information of the training sample refers to a label vector for a target service, and for example, in an insurance service, corresponding label information may refer to a label vector added with corpus information corresponding to insurance services such as heavy insurance purchase, medical insurance withdrawal and the like. And performing vector conversion on the training sample data by adopting the initial vector conversion model to obtain a predicted characteristic vector corresponding to the training sample data, determining a loss value corresponding to the initial vector conversion model according to the predicted characteristic vector and the label vector, and adjusting the initial vector conversion model according to the loss value to obtain the vector conversion model in the embodiment of the application.

The method includes the steps of verifying whether a loss value corresponding to an initial vector conversion model meets a convergence condition or not according to the loss value corresponding to the initial vector conversion model, determining a loss degree to which the loss value of the initial vector conversion model belongs if the loss value does not meet the convergence condition, and adjusting parameters in the initial vector conversion model according to the loss degree to which the loss value belongs. The convergence condition may refer to whether a loss value of the initial vector conversion model is smaller than a preset loss threshold, or whether the iteration number of the initial vector conversion model reaches a preset iteration number. If the loss value of the initial vector conversion model is smaller than the preset loss threshold, it is determined that the initial vector conversion model meets the convergence condition, and the initial vector conversion model with the loss value smaller than the preset loss threshold is determined as the vector conversion model in the embodiment of the present application. If the loss value of the initial vector conversion model is larger than or equal to the preset loss threshold value, determining that the initial vector conversion model does not meet the convergence condition, and continuing training the initial vector conversion model; or if the iteration times of the initial vector conversion model are greater than the preset iteration times, determining that the initial vector conversion model meets the convergence condition, and determining the initial vector conversion model with the iteration times greater than the preset iteration times as the vector conversion model in the embodiment of the application; and if the iteration times of the initial vector conversion model are less than or equal to the preset iteration times, determining that the initial vector conversion model does not meet the convergence condition, and continuing to train the initial vector conversion model.

The loss function corresponding to the initial vector conversion model can be expressed by the following formula (1).

loss_total＝λloss_base+(1-λ)loss_insure (1)

Therein, loss_baseRepresents the loss, of the initial vector transformation model (i.e., the base BERT model)_insureRepresenting the loss of the current corpus environment and lambda represents the hyper-parameter.

Optionally, the specific manner of obtaining the intention probabilities between the information feature vector and the N candidate intentions by the second device may include: and the second equipment inputs the information characteristic vector to a neural network layer in the intention recognition model, and obtains a prediction vector corresponding to the information characteristic vector according to a weight matrix corresponding to the neural network layer. The second device determines intent probabilities between the information feature vectors and the N candidate intentions, respectively, in a classifier of the intent recognition model.

After the second device obtains the information feature vector corresponding to the service consultation information, the information feature vector can be input to a neural network layer in the intention recognition model, and a prediction vector corresponding to the information feature vector is obtained according to a corresponding weight matrix in the neural network layer. The neural network layer in the intention recognition model can simulate the human brain, and performs intention prediction on the information characteristic vector corresponding to the business consultation information to obtain a prediction vector corresponding to the information characteristic vector. After the second device obtains the prediction vectors corresponding to the information feature vectors, the intention probabilities between the prediction vectors and the N candidate intentions respectively can be determined in a classifier of the intention recognition model.

Wherein the N candidate intentions are associated with the target business, such as in insurance business, the N candidate intentions may refer to "heavy insurance purchase", "vehicle insurance refund", "accident insurance claim", "medical insurance consultation", and so on, which are associated with the insurance business. The method comprises the steps of obtaining an intention recognition model during training, obtaining an initial intention recognition model, intention training sample data and marking intention information corresponding to the intention training sample data, adopting the initial intention recognition model to perform intention recognition on the intention training sample data to obtain a predicted intention probability corresponding to the intention training sample data, determining a loss value of the initial intention recognition model according to the predicted intention probability and the marking intention information, adjusting network parameters in the initial intention recognition model according to the loss value corresponding to the initial intention recognition model, and determining the initial intention recognition model meeting a convergence condition to be the intention recognition model when the initial intention recognition model meets the convergence condition.

As shown in fig. 3, fig. 3 is a schematic diagram of an intention recognition model provided in an embodiment of the present application, and as shown in fig. 3, after an information feature vector corresponding to service consultation information is obtained, the information feature vector may be input into a neural network layer in the intention recognition model, and the number of the neural network layers in the intention recognition model may be multiple layers. Each layer of Neural Network layer comprises a perception layer, the Neural Network layer in the intention recognition model refers to an Artificial Neural Network (Artificial Neural Network), abstracting a human brain neuron Network from the information processing angle, so as to establish a simple model, forming different networks according to different connection modes, and the Neural Network layer is formed by connecting a large number of nodes (or called neurons) with each other. And carrying out multiple nonlinear combination on the information characteristic vector corresponding to the service consultation information through a perception layer in the neural network layer to obtain a prediction vector corresponding to the information characteristic vector. After the prediction vectors corresponding to the information characteristic vectors are obtained, the prediction vectors corresponding to the information characteristic vectors are classified through a classifier in the intention recognition model, and intention probabilities between the information characteristic vectors and the N candidate intentions are obtained.

Fig. 4 is a structural diagram of a sensing layer in a neural network layer according to an embodiment of the present invention, and fig. 4 shows a sensing layer neuron in the neural network layer, which is a basic unit of the sensing layer in the neural network, as shown in fig. 4. As shown in fig. 4, the sensing neuron may receive an input information feature vector x, and perform nonlinear combination through a sensing layer in a plurality of neural network layers to obtain a prediction vector corresponding to the information feature vector. Perceptual neuron pair information feature vector q, bias coefficient (i.e., external bias) b, and weight coefficient ω_jAnd carrying out nonlinear combination and obtaining an output result y through an activation function. Wherein, for information characteristic vector q, bias coefficient b and weight coefficient omega_jThe nonlinear combination can be expressed by equation (2).

C＝qω_j+b (2)

Wherein C in the formula (2) refers to the information feature vector x, the offset coefficient B, and the weight coefficient ω_jA value obtained by nonlinear combination, q is an information feature vector, omega_jThe weight coefficient of the jth perception layer is referred to, b is a bias coefficient, and j is a perception layer.

Obtaining an information characteristic vector q, an offset coefficient b and a weight coefficient omega_jAfter the value C obtained by nonlinear combination is carried out, the output value of the j-th perception layer neural unit can be obtained through an activation function. The function expression corresponding to the activation function can be expressed by formula (3).

y＝a(C) (3)

Wherein y in the formula (3) refers to the output value of the j-th perception layer neural unit, and a refers to the activation function.

Therefore, the output value y corresponding to the sensing layer in each neural network is obtained through the method described above, and finally, the output value x (i.e., the prediction vector corresponding to the information feature vector) of the multilayer neural network is obtained according to the output value y corresponding to the sensing layer in each neural network.

In the classifier of the intention recognition model, the following formula (4) may be adopted to determine the intention probabilities between the prediction vectors and the N candidate intentions, respectively.

Wherein i in the formula (4) is an intention, x is a prediction vector corresponding to the information feature vector, and W is_iA classification weight coefficient referring to intention i, c referring to the overall intention, p (y)_i| x) represents the probability of belonging to the information feature vector to the intention i.

After obtaining the intention probabilities between the information feature vector and the N candidate intentions, the second device may determine target intentions from the N candidate intentions according to the intention probabilities, where the number of the target intentions may be 1 or multiple, and the present application does not limit this. If the candidate intention corresponding to the highest intention probability may be determined as the target intention, the candidate intentions corresponding to the first three intention probabilities may be determined as the target intention after sorting the intention probabilities in descending order.

After the second device obtains the target intention corresponding to the service consultation information, M target material data corresponding to the target intention can be obtained from Q candidate material data in the material index library associated with the target service according to the target intention, wherein Q is a positive integer. Specifically, the second device may pre-establish a material index library, add an intention tag to each of the Q candidate material data, and establish a correspondence between the candidate material data and the candidate intention. The candidate material data may be in the form of text material data, video material data, picture material data, and the like. As in insurance business, the material data corresponding to the detailed process of purchasing of the heavy insurance can add an intention label of purchasing the heavy insurance. After a material index library corresponding to the candidate intents and the candidate material data is established, M target material data corresponding to the target intents can be determined from the material index library according to the target intents. Wherein, considering the actual performance influence, the number of M target material data recalled from the material index library can be controlled to be about 50-100. The candidate material data in the material index library can be standardized material data with normalization and specialization collected in advance, so that the normalization and specialization of service data recommendation can be improved, the service image and the service level of the target service can be improved, and the service quality can be improved.

Optionally, the specific manner of displaying the M pieces of target material data by the second device may include: and the second equipment acquires material attribute characteristics corresponding to the M pieces of target material data respectively, and acquires target user attribute characteristics corresponding to target users providing the business consultation information. And the second equipment determines recommended evaluation values corresponding to the M target material data respectively according to the material attribute characteristics, the target user attribute characteristics and the information characteristic vectors corresponding to the service consultation information, sorts the M recommended evaluation values, and displays the M target material data according to the sorted M recommended evaluation values.

Specifically, after obtaining M target material data corresponding to the target intention from the material index library, the second device may obtain material attribute features corresponding to the M target material data, respectively, and obtain user attribute features corresponding to the target user providing the service consultation information. The second device may determine, according to the material attribute feature, the target user attribute feature, and the information feature vector corresponding to the business consultation information corresponding to each target material data of the M target material data, the recommended evaluation value corresponding to each target material data of the M target material data, that is, each target material data of the M target material data is scored in combination with the material attribute feature, the user attribute feature, and the information feature vector. After obtaining the recommended evaluation values corresponding to the M candidate target material data, the second device may rank the M recommended evaluation values, and display the M target material data according to the ranked M recommended evaluation values. For example, after the M recommended evaluation values are sorted in a descending sorting manner, the target material data corresponding to the recommended evaluation value sorted in the front can be displayed.

Optionally, the material attribute features include material semantic features and material type features, and the specific manner of determining, by the second device, the recommended evaluation values corresponding to the M target material data according to the material attribute features, the user attribute features, and the information feature vectors corresponding to the service consultation information may include: and the second equipment inputs the material semantic features, the user attribute features and the information feature vectors corresponding to the service consultation information into the recommendation model, and determines the prediction evaluation values corresponding to the M target material data respectively according to the recommendation model. And the second equipment acquires weight coefficients corresponding to the M target material data respectively according to the material type characteristics, and determines recommended evaluation values corresponding to the M target material data respectively according to the weight coefficients and the predicted evaluation values.

Specifically, the second device may input the material semantic features, the user attribute features, and the information feature vectors corresponding to the service consultation information to the recommendation model, and determine the prediction evaluation values corresponding to the M pieces of target material data, respectively, according to the recommendation model. The material semantic features corresponding to each target material data refer to semantic features corresponding to each target material data, such as keywords in each target material data or target products corresponding to each target material data, and the like. The user attribute features corresponding to the target user refer to the age, gender, behavior preference and the like of the target user, and recommendation evaluation is performed on each target material data by combining the material semantic features, the user attribute features and the information feature vectors to obtain a recommendation evaluation value corresponding to each target material data, namely the association degree of each target material data and the target user is determined, and the interest degree of the target user corresponding to each target material data is determined.

The recommendation model in the embodiment of the present application may be an FM-FTRL (fan Machine-based, the customized Leader, online Machine learning algorithm model) recommendation model, where the FM-FTRL recommendation model is an FM model (Factor Machine, a factorization model, and may be used to learn an interactive hidden relationship between features), and the FM-FTRL recommendation model may be used to learn a cross hidden relationship between material semantic features, user attribute features, and information feature vectors corresponding to service consultation information, so as to obtain a recommendation evaluation value corresponding to each target material data. The functional expression of the FM-FTRL recommendation model can be expressed by the following formula (5).

Wherein w in the formula (5) is a parameter of the input feature, H_iThe method comprises the steps of inputting a feature vector sum (namely a total vector obtained by splicing learning material semantic features, user attribute features and information feature vectors corresponding to business consultation information), V is a hidden vector, each feature vector corresponds to a hidden vector, M is a feature total number, and h is each feature vector (for example, any one of the learning material semantic features or the user attribute features or the information feature vectors corresponding to the business consultation information).<w,H_i>Is linear impact, can obtain independent characteristic information corresponding to each characteristic vector,

the method refers to linear regression, and combined feature information corresponding to a plurality of feature vectors can be obtained. The second device can also collect feedback information of a target user for target material data to update and iterate the FM-FTRL recommendation model in the application, so that the real-time response capability of the FM-FTRL recommendation model is improved, and meanwhile, the requirements of increasing material/cold start exposure in real time and the like can be met. Obtaining each target material data pair through FM-FTRL recommendation modelAnd (4) the corresponding prediction evaluation value is to grade each target material data so as to provide reference information for recommending and consulting feedback information for the target user.

The second device may further obtain weight coefficients corresponding to the M target material data according to the material type feature corresponding to each target material data, that is, the weight coefficient corresponding to each candidate material data may be determined in advance according to the material type feature corresponding to each candidate material data. The material type characteristic corresponding to the material data may refer to importance or exposure rate of a product corresponding to the material data, for example, the more important product may have a higher weight coefficient, and the more product may have a higher weight coefficient. If the candidate material data K is material data corresponding to a new product mainly promoted at the current stage, the weight coefficient of the candidate material data K may be set higher.

After the second device obtains the weight coefficients and the prediction evaluation values corresponding to the M target material data, the recommended evaluation value corresponding to each target material data in the M target material data can be obtained according to the weight coefficient and the prediction evaluation value corresponding to each target material data. If the weight coefficient corresponding to the target material data S is e, and the predicted evaluation value corresponding to the target material data S is z, the recommended evaluation value corresponding to the target material data S may be a product of the weight coefficient and the predicted evaluation value, i.e., e × z. After obtaining the recommended evaluation value corresponding to each target material data, the second device may sort the M recommended evaluation values to obtain the sorted M recommended evaluation values, and display the M target material data according to the sorted M recommended evaluation values. For example, the M target material data may be sorted according to the sorted M recommended evaluation values, for example, the M recommended evaluation values may be sorted in a descending order, the M recommended evaluation values may be sorted in an ascending order, and the like.

After the second device obtains the prediction evaluation value corresponding to each target material data, the second device can sort the prediction evaluation values corresponding to each target material data to obtain M sorted prediction evaluation values, and sort the M target material data according to the M sorted prediction evaluation values to obtain M sorted target material data. Specifically, the target material data in each target intention can be sorted, the sorted target material data can be adjusted after the sorted target material data in each target intention is obtained, for example, the material type characteristics of each target material data are obtained to adjust the sorted target material data, for example, the material type characteristics can include a video material type, a text material type and a picture material type, a certain amount of target material data are reserved under the video material type, a certain amount of target material data are reserved under the text material type, a certain amount of target material data are reserved under the picture material type, namely, a certain amount of target material data are reserved under each type of material, so that different potential requirements of target users are met, and meanwhile, the storage space can be saved. And the reserved target material data can be determined according to the prediction evaluation value corresponding to each target material data, and the reserved target material characteristics are displayed in the interactive interface corresponding to the second equipment. The M target material data are configured and optimized by using a certain configuration rule, so that the M target material data meet the requirements of target services as much as possible. The M target material data are more organized, so that consultation feedback information can be provided for the target user subsequently, and service data recommendation efficiency is provided.

For example, in the insurance business, if the target intention is to purchase a heavy insurance, after 60 items of labeled material data are recalled from the material index library according to the target intention of purchasing the heavy insurance, the corresponding prediction evaluation values of the 60 items of labeled material data can be obtained by using the recommendation model. And then, acquiring material type characteristics corresponding to the 60-item label material data, and classifying the 60-item label material data, namely dividing the 60-item label material data into video material data, character material data or picture material data. And sequencing the target material data under each material type according to the prediction evaluation value corresponding to the 60-item target material data respectively, sequencing the 60-item target material data, and if the target material data under the video material type is 20, sequencing the 20-item target material data under the video material type according to the prediction evaluation value corresponding to the 20-item target material data, and then retaining the previously sequenced target material data. And displaying the M target material data on an interactive interface corresponding to the second device.

And S103, the second equipment responds to the selection operation aiming at the M target material data, and the target material data determined by the selection operation is determined as the consultation feedback information in the M target material data.

Specifically, after the M pieces of target material data are displayed on the interactive interface corresponding to the second device, the service staff may select the advisory feedback information to be sent to the target user in the interactive interface where the M pieces of target material data are displayed in the second device, where the service staff refers to the staff corresponding to the target service. The second device may respond to a selection operation of the business person for each of the M target material data, and determine the target material data determined by the selection operation as the consultation feedback information among the M target material data. Because the second device directly determines the feedback information for feeding back to the target user from the M target material data, the error rate is high, and the consultation is more mechanized, so that more accurate consultation foul information cannot be well provided for the target user. Therefore, the scheme can display the M pieces of target material data on a display interface (namely an interactive interface) of the second equipment where the service personnel are located, the service personnel determine the consultation feedback information finally sent to the target user, the accuracy of service data recommendation can be provided, better consultation experience is brought to the target user, and the user problem is better solved.

And S104, the second equipment sends the consultation feedback information to the first equipment.

Specifically, after obtaining the advisory feedback information, the second device may send the advisory feedback information to the first device where the target user is located. And after the first equipment receives the consultation feedback information sent by the second equipment, displaying the consultation feedback information on a display interface corresponding to the first equipment so as to answer the business consultation information corresponding to the target user.

As shown in fig. 5, fig. 5 is a schematic diagram illustrating determination of advisory feedback information corresponding to business advisory information provided in an embodiment of the present application, and as shown in fig. 5, after a target user inputs business advisory information for a target business on a display interface corresponding to a first device where the target user is located, the first device sends the business advisory information for the target business to a second device. After receiving the service consultation information sent by the first equipment, the second equipment extracts keywords 50a from the service consultation information, and extracts consultation keywords in the service consultation information. The second device performs feature vector conversion 50b on the consultation keywords in the service consultation information to obtain information feature vectors corresponding to the consultation keywords. After the information feature vector corresponding to the consultation keyword is obtained, the second device performs intention recognition 50c on the business consultation information according to the information feature vector to obtain a target intention corresponding to the business consultation information. And the second equipment recalls the material from the material index library for 50d according to the target intention corresponding to the business consultation information to obtain M target material data corresponding to the target intention. The second device may obtain predicted evaluation values corresponding to the M target material data respectively according to the user attribute characteristics of the target user, the material attribute characteristics corresponding to each target material data, and the information feature vector corresponding to the service consultation information, and perform material sorting 50e on the M target material data according to the predicted evaluation values corresponding to the M target material data respectively, to obtain M sorted target material data. The second device may adjust the M target materials according to a certain configuration policy (e.g., determining a weight coefficient according to material type characteristics corresponding to the target material data, and then obtaining a recommended evaluation value according to the weight coefficient and the predicted evaluation value). And displaying the adjusted M target material data on a display interface corresponding to the second equipment, and determining consultation feedback information from the displayed M target material data by the service personnel.

The embodiment of the application can also be applied to the case that after the target user triggers the determination operation for the manual reply, the first device where the target user is located can be connected with the communication channel between the second device, so that the communication channel between the target user and the service staff is connected. And after receiving the service consultation information of the target user aiming at the target service, the first equipment sends the service consultation information to the second equipment. And after receiving the service consultation information sent by the first equipment, the second equipment determines consultation feedback information corresponding to the service consultation information and returns the consultation feedback information to the first equipment. After the first equipment receives the consultation feedback information sent by the second equipment, the consultation feedback information is displayed on the display interface corresponding to the first equipment, so that better consultation experience is brought to a target user and the user problem is better solved. Of course, the embodiment of the present application may also be applied to that after the target user inputs the service consultation information for the target service, the first device directly sends the service consultation information to the second device, receives the consultation feedback information corresponding to the service consultation information returned by the second device, and displays the consultation feedback information.

As shown in fig. 6, fig. 6 is a schematic view illustrating that the consultation feedback information for the insurance service is recommended to the target user according to the embodiment of the present application, and as shown in fig. 6, if the target user wants to obtain the operation flow information about how to operate the payment renewal of the heavy insurance service, the target user may input the service consultation information for the heavy insurance service on the corresponding interactive page in the first device 60a (e.g., a mobile phone). After receiving the heavy risk consultation information of the target user for the heavy risk, the first device 60a transmits the heavy risk consultation information to the second device 60b, and the second device 60b may transmit the heavy risk consultation information to the server 60c corresponding to the second device 60 b. After the server 60c corresponding to the second device 60b receives the heavy disease consultation information of the target user for the heavy disease, the server 60c may obtain an information feature vector corresponding to the heavy disease consultation information, determine a target intention corresponding to the heavy disease consultation information according to the information feature vector, and recall M target material data associated with the target intention from the material index library. The server 60c transmits the M target material data corresponding to the critical illness consultation information to the second device 60b, and the second device 60b displays the M target material data on the corresponding interactive interface. The service personnel can determine the heavy disease consultation feedback information fed back to the target user from the M target material data, the second equipment 60b responds to the selection operation of the service personnel on the M target material data, and the target material data determined by the selection operation is determined as the heavy disease consultation feedback information from the M target material data. The second device 60b returns the critical illness consultation feedback information to the first device 60a, and after receiving the critical illness consultation information returned by the second device 60b, the first device 60a displays the critical illness consultation information on a display interface corresponding to the first device 60a, and the target user can view the critical illness consultation information on a corresponding page.

As shown in fig. 7, fig. 7 is a schematic view of a method for determining consultation feedback information in a related technology provided in an embodiment of the present application, and as shown in fig. 7, in the related technology, after a service person (such as a customer service person or a service manager) corresponding to a target service receives service consultation information of the target service for the target service, the service person needs to query material data associated with the service consultation information from a terminal device where the service person is located according to the service consultation information sent by the target user, or send the material data associated with screenshot from some web pages or some applets to the target user. As shown in fig. 7, in the related art, after the service staff receives the service consultation information sent by the target user, reference material data for reference of the service staff is not provided on the terminal display interface corresponding to the service staff. The business personnel need to search the consultation feedback information corresponding to the business consultation information by themselves, so that the obtained consultation feedback information is not standard, and for the same business consultation information, the consultation feedback information found by each business personnel can be different, so that the professional and uniformity of the business service are caused to have problems, and even the service image of the target business can be influenced. Meanwhile, business personnel search for the consultation feedback information by themselves, so that the waiting time of the target user is too long, and the experience of the target user is poor. The service personnel can not give consideration to more user consultation, so that the manpower efficiency is low, the sending error is easy to occur, the misoperation is generated, and the accuracy and the efficiency of service data recommendation are low.

As shown in fig. 8a, fig. 8a is a schematic view determined to display target material data according to an embodiment of the present application, where after receiving insurance service consultation information "how to pay for the renewal" of the target user for the insurance service sent by the first device, the second device obtains an information feature vector corresponding to the insurance service consultation information "how to pay for the renewal". And the second equipment determines a corresponding target intention of insurance service consultation information, namely 'how to pay for the next period' according to the information characteristic vector, recalls M target material data associated with the target intention from the material index library, and displays the M target material data on a corresponding display interface. As shown in fig. 8a, the service staff can see the insurance service consultation information "what is due for the renewal of the fee" sent by the target user for the insurance service on the display interface corresponding to the second device. As shown in fig. 8a, the right end of the display interface corresponding to the second device displays M target material data corresponding to insurance business consultation information, that is, M answers for answering "how to pay for the next due" consultation information, and the business personnel can determine insurance consultation feedback information sent to the target user from the M target material data displayed by the second device. As shown in fig. 8a, when displaying M target material data, the target material data may be displayed according to a material type corresponding to the target material data, for example, the target material data of a text type is displayed first, and then the target material data of a picture type is displayed, which is convenient for business personnel to select.

As shown in fig. 8b, fig. 8b is a schematic diagram of determining advisory feedback information provided in an embodiment of the present application, and as shown in fig. 8b, a service person may determine insurance advisory feedback information to be sent to a target user from text-type target material data, and after determining the target material data to be sent to the target user, the service person may directly click a "send" touch key, and a second device may determine corresponding target material data as advisory feedback information and send the advisory feedback information to a first device where the target user is located. As shown in fig. 8b, the service person may also click a "copy" touch key, paste the copy in the corresponding sending window, and send the copy, so that the service person may adjust the copied target material data, and may also send the copied target material data to other users, thereby improving the convenience of the service person in operation and simultaneously improving the work efficiency.

As shown in fig. 8c, fig. 8c is a schematic diagram of determining advisory feedback information according to an embodiment of the present application, and as shown in fig. 8c, a service person may determine target material data of a picture type to be sent to a target user in a corresponding picture template (i.e., target material data in the picture type). Similarly, the service personnel can click to send directly or copy before sending.

As shown in fig. 8d, fig. 8d is a schematic view of target material data provided in the embodiment of the present application, and as shown in fig. 8d, for example, in the insurance service, for the insurance consultation information of "what is to be paid for renewal" sent by the target user, the corresponding target material data may be a specific operation process of the payment for renewal. The method comprises the following steps: selecting an insurance type; the second step is that: selecting a specific insurance; the third step: selecting a guarantee amount; the fourth step: selecting the payment years and other renewal payment specific processes.

The embodiment of the application can also be applied to medical consultation services, and if a target user wants to consult medical consultation information such as medical insurance information or a visiting process for an oral department, the corresponding medical consultation information can be input in a corresponding page through a mobile phone. After receiving the medical consultation information, the terminal equipment corresponding to the business personnel can determine M target material data corresponding to the medical consultation information, and the corresponding business personnel can select the target material data to be sent to the target user from the M target material data. And the terminal equipment corresponding to the service personnel responds to the determination operation of the service personnel on the M pieces of target material data, after the medical feedback information is determined, the medical feedback information is sent to the terminal equipment where the target user is located, the terminal equipment where the target user is located displays the medical feedback information, and answers the medical feedback information corresponding to the target user.

In the embodiment of the application, the first device responds to the information input operation aiming at the target service, and the service consultation information determined by the information input operation is sent to the second device. And the second equipment acquires M target material data matched with the service consultation information in the target service and displays the M target material data, wherein M is a positive integer. The second equipment responds to selection operation aiming at the M target material data, determines the target material data determined by the selection operation as consultation feedback information in the M target material data, and sends the consultation feedback information to the first equipment. It can be seen that, for the service consultation information sent by the first device, the first and second devices can preliminarily screen M target material data matched with the service consultation information in the target service, and display the M target material data, consultation feedback information corresponding to the service consultation information can be determined from the displayed M target material data, and the consultation feedback information is returned to the first device as reply information of the service consultation information, so that tedious manual operation can be reduced, the recommendation efficiency of the service data can be improved, and the accuracy of service data recommendation can be improved by using the consultation feedback information determined from the preliminarily screened M target material data. Meanwhile, the method and the system can improve convenience and per-capita productivity of business personnel, improve standardization and speciality of business data recommendation, bring better consultation experience to the user, better solve the problem of the user and improve service experience corresponding to the target business.

As shown in fig. 9, fig. 9 is a schematic diagram of a service data recommendation method provided in an embodiment of the present application, where the method may be executed by a computer device, and the method may be executed by a computer device, where the computer device may be a server (such as the server 10 in fig. 1), or a target user terminal (such as any one of the target user terminals in the target user terminal cluster in fig. 1), or a system composed of the server and the target user terminal, which is not limited in this application. As shown in fig. 9, the steps of the service data recommendation method include S201-202.

S201, the first equipment responds to information input operation aiming at the target service, and sends the service consultation information determined by the information input operation to the second equipment so that the second equipment determines consultation feedback information corresponding to the service consultation information in M target material data contained in the target service; m is a positive integer.

Specifically, if the target user wants to consult the target service, the target user may input service consultation information for the target service in the interactive page corresponding to the first device. The first device may respond to an information input operation of a target user for a target service, send service consultation information determined by the information input operation, and send service consultation information for the target service to the second device. After receiving the service consultation information sent by the first device, the second device may obtain an information feature vector corresponding to the service consultation information, determine a target intention corresponding to the service consultation information according to the information feature vector, recall M target material data associated with the target intention from a material index library corresponding to the target service, and determine consultation feedback information corresponding to the service consultation information in the M target material data.

S202, the first equipment receives the consultation feedback information sent by the second equipment and displays the consultation feedback information.

Specifically, after determining the advisory feedback information corresponding to the service advisory information, the second device returns the advisory feedback information to the first device. The first device can receive the consultation feedback information sent by the second device and display the consultation feedback information on the interactive interface corresponding to the first device, so that the target user can view the consultation feedback information and provide answers for the target user.

As shown in fig. 10, fig. 10 is a schematic view illustrating displaying of consultation feedback information provided in an embodiment of the present application, and as shown in fig. 10, the interactive interface shown in fig. 10 is an interactive interface of a target user side, and the target user may input service consultation information for a target service in the interactive interface corresponding to the first device. For example, in the insurance service, when the target user wants to consult a specific operation flow of insurance renewal payment, the service consultation information "how to do renewal payment" for the insurance service can be input in the customer service interaction interface corresponding to the insurance service found in the first device (e.g., mobile phone) corresponding to the target user. After acquiring the "how do the renewal payment" input by the target user, the first device corresponding to the target user may send the "how do the renewal payment" to the second device corresponding to the insurance service, and after receiving the "how do the renewal payment" the second device determines the consultation feedback information corresponding to the "how do the renewal payment", such as a text type renewal payment operation flow or a picture type renewal payment flow. And the corresponding consultation feedback information of 'how to pay for the next term' is returned to the first equipment, and the first equipment displays the consultation feedback information on the corresponding interactive interface.

Optionally, the first device may further obtain user behavior data for the advisory feedback information, and send the user behavior data to the second device, so that the second device updates the recommendation model based on the user behavior data. The recommendation model is used for determining the prediction evaluation values corresponding to the M target material data, and the prediction evaluation values provide basis conditions for the second equipment to determine the consultation feedback information.

Specifically, after the first device displays the consultation feedback information on the corresponding interactive interface, the first device may further obtain user behavior data of the target user for the consultation feedback information, where the user behavior data may be: and in a preset time period, the target user checks the consultation feedback information, or the target user does not check the consultation feedback information. After obtaining the user behavior data corresponding to the target user, the first device may send the user behavior data to the second device. After receiving the user behavior data aiming at the consultation feedback information sent by the first equipment, the second equipment can set feedback label information for the target material data corresponding to the consultation feedback information according to the user behavior data. If the user behavior data indicates that the user has viewed the advisory feedback information, the feedback tag information of the target material data corresponding to the advisory feedback information may be set to "1", and if the user behavior data indicates that the user has not viewed the advisory feedback information, the feedback tag information of the target material data corresponding to the advisory feedback information may be set to "0". And updating the recommendation model according to the target material data corresponding to the consultation feedback information and the corresponding feedback label information, so that the material recommendation capability of the recommendation model can be provided, the real-time response capability of the recommendation model can be improved, and the iteration efficiency of the recommendation model can be accelerated.

In the embodiment of the application, the first device receives the consultation feedback information sent by the second device, and displays the consultation feedback information, so that the second device determines the consultation feedback information corresponding to the service consultation information in the M target material data contained in the target service. The first device may receive the advisory feedback information sent by the second device and display the advisory feedback information. The consultation feedback information is sent to the second equipment through the first equipment, so that the second equipment determines the consultation feedback information corresponding to the business consultation information in the M target material data contained in the target business, and the accuracy and the efficiency of recommending the business data can be improved. Meanwhile, the user behavior data aiming at the consultation feedback information can be sent to the second equipment through the first equipment, so that the second equipment updates the recommendation model based on the user behavior data, the accuracy of the recommendation model in the second equipment is improved, and the accuracy of service data recommendation is further improved.

Referring to fig. 11, fig. 11 is a schematic structural diagram of a service data recommendation device according to an embodiment of the present application. The service data recommendation apparatus 1 may be a computer program (including program code) running in a computer device, for example, the service data recommendation apparatus 1 is an application software; the device 1 may be configured to perform corresponding steps in the service data recommendation method provided in the embodiment of the present application. As shown in fig. 11, the service data recommendation apparatus 1 may include: the device comprises a first determination module 11, a display module 12, a second determination module 13 and a sending module 14.

A first sending module 11, configured to respond to an information input operation for a target service, and send service consultation information determined by the information input operation to a second device;

the display module 12 is configured to obtain M pieces of target material data matched with the service consultation information in the target service, and display the M pieces of target material data; m is a positive integer;

a first determining module 13, configured to respond to a selection operation for the M pieces of target material data, and determine, among the M pieces of target material data, the target material data determined by the selection operation as consultation feedback information;

and a second sending module 14, configured to send the advisory feedback information to the first device.

Wherein the first determining module 13 includes:

a first obtaining unit 1301, configured to receive service consultation information sent by a first device, and obtain an information feature vector corresponding to the service consultation information;

a second obtaining unit 1302, configured to obtain intention probabilities between the information feature vectors and the N candidate intentions, respectively; n is a positive integer;

a first determining unit 1303, configured to determine, according to the intention probability, a target intention corresponding to the service consultation information from the N candidate intentions;

a third obtaining unit 1304, configured to obtain M pieces of target material data that match the target intention in the target service.

The first obtaining unit 1301 is specifically configured to:

The second obtaining unit 1302 is specifically configured to:

Wherein, the display module 12 includes:

a fourth obtaining unit 1201, configured to obtain material attribute features corresponding to the M pieces of target material data, and obtain a user attribute feature corresponding to a target user providing service consulting information;

a second determining unit 1202, configured to determine, according to the material attribute features, the user attribute features, and the information feature vectors corresponding to the service consultation information, recommended evaluation values corresponding to the M pieces of target material data, respectively;

the display unit 1203 is configured to sort the M recommended evaluation values, and display the M target material data according to the sorted M recommended evaluation values.

the second determining unit 1202 is specifically configured to:

According to an embodiment of the present application, the steps involved in the service data recommendation method shown in fig. 2 may be performed by various modules in the service data recommendation device 2 shown in fig. 11. For example, step S101 shown in fig. 2 may be performed by the first transmitting module 11 in fig. 11, and step S102 shown in fig. 2 may be performed by the display module 12 in fig. 11; step S103 shown in fig. 2 may be performed by the first determination module 13 in fig. 11; step S104 shown in fig. 2 may be performed by the second sending module 14 in fig. 11.

According to an embodiment of the present application, each module in the service data recommendation apparatus 1 shown in fig. 11 may be respectively or entirely combined into one or several units to form the unit, or some unit(s) may be further split into multiple sub-units with smaller functions, which may implement the same operation without affecting implementation of technical effects of the embodiment of the present application. The modules are divided based on logic functions, and in practical application, the functions of one module can be realized by a plurality of units, or the functions of a plurality of modules can be realized by one unit. In other embodiments of the present application, the testing device may also include other units, and in practical applications, these functions may also be implemented by the assistance of other units, and may be implemented by cooperation of a plurality of units.

In the embodiment of the application, the first device responds to the information input operation aiming at the target service, and the service consultation information determined by the information input operation is sent to the second device. And the second equipment acquires M target material data matched with the service consultation information in the target service and displays the M target material data, wherein M is a positive integer. The M target material data matched with the business consultation information are obtained in the target business, reference information can be provided for business personnel to determine the consultation feedback information in the M target material data, and the standardization and the efficiency of business data recommendation can be improved. The second equipment responds to selection operation aiming at the M target material data, determines the target material data determined by the selection operation as consultation feedback information in the M target material data, and sends the consultation feedback information to the first equipment. It can be seen that, for the service consultation information sent by the first device, the first and second devices can preliminarily screen M target material data matched with the service consultation information in the target service, and display the M target material data, consultation feedback information corresponding to the service consultation information can be determined from the displayed M target material data, and the consultation feedback information is returned to the first device as reply information of the service consultation information, so that tedious manual operation can be reduced, the recommendation efficiency of the service data can be improved, and the accuracy of service data recommendation can be improved by using the consultation feedback information determined from the preliminarily screened M target material data. Meanwhile, the method and the system can improve convenience and per-capita productivity of business personnel, improve standardization and speciality of business data recommendation, bring better consultation experience to the user, better solve the problem of the user and improve service experience corresponding to the target business.

Referring to fig. 12, fig. 12 is a schematic structural diagram of a service data recommendation device according to an embodiment of the present application. The service data recommendation device 2 may be a computer program (including program code) running in a computer device, for example, the service data recommendation device 2 is an application software; the device 2 may be configured to perform corresponding steps in the service data recommendation method provided in the embodiment of the present application. As shown in fig. 12, the service data recommendation apparatus 2 may include: a third determining module 21, a receiving module 22 and a returning module 23.

The second determining module 21 is configured to respond to an information input operation for the target service, and send the service consultation information determined by the information input operation to the second device, so that the second device determines consultation feedback information corresponding to the service consultation information in the M pieces of target material data included in the target service; m is a positive integer;

and the receiving module 22 is configured to receive the advisory feedback information sent by the second device and display the advisory feedback information.

Wherein, the device 2 further comprises:

the third sending module 23 is configured to obtain user behavior data for the advisory feedback information, and send the user behavior data to the second device, so that the second device updates the recommendation model based on the user behavior data; the recommendation model is used for determining the prediction evaluation values corresponding to the M target material data, and the prediction evaluation values provide basis conditions for the second equipment to determine the consultation feedback information.

According to an embodiment of the present application, the steps involved in the service data recommendation method shown in fig. 9 may be performed by the respective modules in the service data recommendation device 2 shown in fig. 12. For example, step S201 shown in fig. 9 may be performed by the second determining module 21 in fig. 12, step S202 shown in fig. 9 may be performed by the receiving module 22 in fig. 12, and so on.

According to an embodiment of the present application, each module in the service data recommendation apparatus 2 shown in fig. 12 may be respectively or entirely combined into one or several units to form the unit, or some unit(s) may be further split into multiple sub-units with smaller functions, which may implement the same operation without affecting implementation of technical effects of the embodiment of the present application. The modules are divided based on logic functions, and in practical application, the functions of one module can be realized by a plurality of units, or the functions of a plurality of modules can be realized by one unit. In other embodiments of the present application, the testing device may also include other units, and in practical applications, these functions may also be implemented by the assistance of other units, and may be implemented by cooperation of a plurality of units.

According to an embodiment of the present application, the service data recommendation apparatus 1 shown in fig. 11 and the service data recommendation apparatus 2 shown in fig. 12 may be constructed by running a computer program (including program codes) capable of executing the steps involved in the corresponding methods shown in fig. 2 or fig. 9 on a general-purpose computer device, such as a computer, including a processing element and a storage element, such as a Central Processing Unit (CPU), a random access storage medium (RAM), a read-only storage medium (ROM), and the like, and implementing the service data recommendation method of the embodiment of the present application. The computer program may be recorded on a computer-readable recording medium, for example, and loaded into and executed by the computing apparatus via the computer-readable recording medium.

Referring to fig. 13, fig. 13 is a schematic structural diagram of a computer device according to an embodiment of the present application. As shown in fig. 13, the computer apparatus 1000 may include: the processor 1001, the network interface 1004, and the memory 1005, and the computer apparatus 1000 may further include: a target user interface 1003, and at least one communication bus 1002. Wherein a communication bus 1002 is used to enable connective communication between these components. The target user interface 1003 may include a Display screen (Display) and a Keyboard (Keyboard), and the selectable target user interface 1003 may also include a standard wired interface and a standard wireless interface. The network interface 1004 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface). The memory 1005 may be a high-speed RAM memory or a non-volatile memory (e.g., at least one disk memory). The memory 1005 may optionally be at least one memory device located remotely from the processor 1001. As shown in fig. 13, a memory 1005, which is a kind of computer-readable storage medium, may include therein an operating system, a network communication module, a target user interface module, and a device control application program.

In the computer device 1000 shown in fig. 13, the network interface 1004 may provide a network communication function; the target user interface 1003 is an interface for providing input to a target user; and the processor 1001 may be used to invoke a device control application stored in the memory 1005 to implement:

optionally, the processor 1001 may be configured to invoke a device control application stored in the memory 1005 to implement:

and sending the consultation feedback information to the first equipment.

responding to information input operation aiming at the target service, and sending the service consultation information determined by the information input operation to the second equipment so that the second equipment determines consultation feedback information corresponding to the service consultation information in M target material data contained in the target service; m is a positive integer;

and receiving the consultation feedback information sent by the second equipment, and displaying the consultation feedback information.

It should be understood that the computer device 1000 described in this embodiment may perform the description of the service data recommendation method in the embodiment corresponding to fig. 2 or fig. 9, and may also perform the description of the service data recommendation device 1 corresponding to fig. 11 and the service data recommendation device 2 corresponding to fig. 12, which is not described herein again. In addition, the beneficial effects of the same method are not described in detail.

According to an aspect of the application, a computer program product or computer program is provided, comprising computer instructions, the computer instructions being stored in a computer readable storage medium. The processor of the computer device reads the computer instruction from the computer-readable storage medium, and the processor executes the computer instruction, so that the computer device can execute the description of the service data recommendation method in the embodiment corresponding to fig. 2 or fig. 9, which is not described herein again. In addition, the beneficial effects of the same method are not described in detail.

By way of example, the program instructions described above may be executed on one computer device, or on multiple computer devices located at one site, or distributed across multiple sites and interconnected by a communication network, which may comprise a blockchain network.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by a computer program, which can be stored in a computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.

The above disclosure is only for the purpose of illustrating the preferred embodiments of the present application and is not to be construed as limiting the scope of the present application, so that the present application is not limited thereto, and all equivalent variations and modifications can be made to the present application.

Claims

1. A business data recommendation method is characterized by comprising the following steps:

the first equipment responds to information input operation aiming at the target service and sends the service consultation information determined by the information input operation to the second equipment;

and the second equipment sends the consultation feedback information to the first equipment.

2. The method as claimed in claim 1, wherein the second device obtains M target material data matched with the service consultation information in the target service, including:

the second equipment receives the service consultation information sent by the first equipment, and obtains an information characteristic vector corresponding to the service consultation information;

the second equipment acquires intention probabilities between the information feature vector and N candidate intentions respectively; n is a positive integer;

3. The method according to claim 2, wherein the obtaining of the information feature vector corresponding to the service consultation information includes:

4. The method of claim 2, wherein the second device obtains intention probabilities between the information feature vector and N candidate intentions, respectively, comprising:

the second equipment inputs the information characteristic vector to a neural network layer in an intention recognition model, and obtains a prediction vector corresponding to the information characteristic vector according to a weight matrix corresponding to the neural network layer;

the second device determines, in a classifier of the intent recognition model, intent probabilities between the prediction vectors and the N candidate intentions, respectively.

5. The method according to claim 1, wherein said displaying said M target material data comprises:

the second equipment acquires material attribute characteristics corresponding to the M pieces of target material data respectively, and acquires user attribute characteristics corresponding to a target user providing the business consultation information;

and the second equipment ranks the M recommended evaluation values and displays the M target material data according to the ranked M recommended evaluation values.

6. The method of claim 5, wherein the material attribute features comprise material semantic features and material type features;

the second device determines, according to the material attribute features, the user attribute features, and the information feature vectors corresponding to the service consultation information, recommended evaluation values corresponding to the M pieces of target material data, respectively, including:

the second equipment inputs the material semantic features, the user attribute features and the information feature vectors corresponding to the service consultation information into a recommendation model, and the prediction evaluation values corresponding to the M pieces of target material data are determined according to the recommendation model;

the second equipment acquires weight coefficients corresponding to the M pieces of target material data respectively according to the material type characteristics;

7. A business data recommendation method is characterized by comprising the following steps:

the method comprises the steps that a first device responds to information input operation aiming at a target service, and service consultation information determined by the information input operation is sent to a second device, so that the second device determines consultation feedback information corresponding to the service consultation information in M target material data contained in the target service; m is a positive integer;

8. The method of claim 7, further comprising:

the first equipment acquires user behavior data aiming at the consultation feedback information and sends the user behavior data to the second equipment so that the second equipment updates a recommendation model based on the user behavior data; the recommendation model is used for determining prediction evaluation values corresponding to the M target material data, and the prediction evaluation values provide basis conditions for the second device to determine the consultation feedback information.

9. A computer device, comprising: a processor and a memory;

the memory stores a computer program which, when executed by the processor, performs the method of any of claims 1 to 7.

10. A computer-readable storage medium, in which a computer program is stored which is adapted to be loaded by a processor and to carry out the method of any one of claims 1 to 7.