CN115601047A

CN115601047A - Data processing method and device

Info

Publication number: CN115601047A
Application number: CN202211164742.9A
Authority: CN
Inventors: 王泽慧
Original assignee: Alibaba China Co Ltd
Current assignee: Alibaba China Co Ltd
Priority date: 2022-09-23
Filing date: 2022-09-23
Publication date: 2023-01-13

Abstract

An embodiment of the present specification provides a data processing method and an apparatus, wherein the data processing method includes: determining an access stream generated by a user for object access, acquiring to-be-processed data corresponding to different access nodes in the access stream, determining an access type corresponding to the user for target object access according to the to-be-processed data, determining to-be-recommended information corresponding to the target access type under the condition that the access type belongs to the target access type, and displaying the to-be-recommended information to the user through an information operation interface.

Description

Data processing method and device

Technical Field

The embodiment of the specification relates to the technical field of computers, in particular to a data processing method.

Background

Poor evaluation rate in the e-commerce field is an important index for platform evaluation of merchants, in practical application, the poor evaluation rate can be divided into aspects of description conformity, logistics service, service attitude and the like, and in the current environment, improvement of the merchants on user experience is often delayed. For example, after the user makes a medium and bad comment on a commodity, the merchant consults the user and optimizes the dissatisfaction of the user, the processing mode is delayed, the processing measures provided by the merchant only play a good and late role, and the stickiness and good sensitivity of the user to the merchant are generally reduced, especially, the situation that the user makes a medium and bad comment due to misunderstanding is more unfortunately; in addition, post-consultation of the user in this manner may also create a new turn of annoyance to the user itself. Meanwhile, the poor and medium comment content of the user often relates to different aspects, and the merchant often faces a large amount of expressions and cannot find the core and concentrated problem to solve, so that the solving efficiency is low.

Therefore, an effective method is needed to solve such problems.

Disclosure of Invention

In view of this, the embodiments of the present specification provide a data processing method. One or more embodiments of the present specification also relate to a data processing apparatus, an information recommendation method, a computing device, a computer-readable storage medium, and a computer program, so as to solve technical deficiencies in the prior art.

According to a first aspect of embodiments herein, there is provided a data processing method including:

determining an access stream generated by object access of a user, and acquiring to-be-processed data corresponding to different access nodes in the access stream;

determining an access type corresponding to the target object access of the user according to the data to be processed;

determining information to be recommended corresponding to the target access type under the condition that the access type belongs to the target access type;

and displaying the information to be recommended to the user through an information operation interface.

According to a second aspect of embodiments herein, there is provided a data processing apparatus comprising:

the acquisition module is configured to determine an access stream generated by object access performed by a user, and acquire to-be-processed data corresponding to different access nodes in the access stream;

the first determining module is configured to determine an access type corresponding to target object access of the user according to the data to be processed;

the second determining module is configured to determine information to be recommended corresponding to the target access type under the condition that the access type belongs to the target access type;

and the display module is configured to display the information to be recommended to the user through an information operation interface.

According to a third aspect of embodiments of the present specification, there is provided an information recommendation method including:

determining an access flow generated by a user for commodity access, and acquiring to-be-processed data corresponding to different access nodes in the access flow;

determining an access type corresponding to target commodity access of the user according to the data to be processed;

and displaying the information to be recommended to the user through a conversation interactive interface.

According to a fourth aspect of embodiments herein, there is provided a computing device comprising:

a memory and a processor;

the memory is used for storing computer-executable instructions, and the processor is used for executing the computer-executable instructions to realize the steps of any one of the data processing method and the information recommendation method.

According to a fourth aspect of embodiments herein, there is provided a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of any one of the data processing method or the information recommendation method.

According to a fifth aspect of embodiments of the present specification, there is provided a computer program, wherein when the computer program is executed in a computer, the computer program causes the computer to execute the steps of the above-described data processing method or the above-described information recommendation method.

In one embodiment of the present specification, an access stream generated by a user performing object access is determined, to-be-processed data corresponding to different access nodes in the access stream is obtained, an access type corresponding to the user performing target object access is determined according to the to-be-processed data, to-be-recommended information corresponding to a target access type is determined when the access type belongs to the target access type, and the to-be-recommended information is presented to the user through an information operation interface.

In the process of object access by a user, the embodiment of the specification predicts the access type of the user by analyzing the dialogue data and/or the access behavior data (to-be-processed data) corresponding to different access nodes in the generated access stream, so that an object provider can intervene the user in advance according to the access type, that is, send information to be recommended to the user, and improve the service experience or the object access satisfaction of the user in the whole object access process through the information to be recommended.

Drawings

FIG. 1 is a schematic diagram of a data processing process provided in one embodiment of the present description;

FIG. 2 is a flow chart of a data processing method provided by an embodiment of the present specification;

FIG. 3a is a flow diagram of a data processing process provided by one embodiment of the present description;

FIG. 3b is a flow diagram of another data processing process provided by one embodiment of the present description;

FIG. 4 is a flowchart illustrating a data processing method according to an embodiment of the present disclosure;

fig. 5 is a schematic structural diagram of a data processing apparatus according to an embodiment of the present specification;

FIG. 6 is a flow chart of a method for recommending information provided in an embodiment of the present specification;

FIG. 7 is a block diagram of an information recommendation system according to an embodiment of the present disclosure;

fig. 8 is a block diagram of a computing device according to an embodiment of the present disclosure.

Detailed Description

In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present specification. This description may be embodied in many different forms and should not be construed as limited to the embodiments set forth herein, as those skilled in the art will be able to make and use the present disclosure without departing from the spirit and scope of the present disclosure.

The terminology used in the description of the one or more embodiments is for the purpose of describing the particular embodiments only and is not intended to be limiting of the description of the one or more embodiments. As used in one or more embodiments of the present specification and the appended claims, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used in one or more embodiments of the present specification refers to and encompasses any and all possible combinations of one or more of the associated listed items.

It will be understood that, although the terms first, second, etc. may be used herein in one or more embodiments to describe various information, these information should not be limited by these terms. These terms are only used to distinguish one type of information from another. For example, a first can also be referred to as a second and, similarly, a second can also be referred to as a first without departing from the scope of one or more embodiments of the present description. The word "if," as used herein, may be interpreted as "at … …" or "at … …" or "in response to a determination," depending on the context.

First, the noun terms to which one or more embodiments of the present specification relate are explained.

multi-head-attribute: the multi-headed attention mechanism, a method of machine learning that maps inputs.

mlp: the multilayer perceptron is a neural network which comprises at least one hidden layer and is composed of fully-connected layers, and the output of each hidden layer is transformed through an activation function.

look-aike: the look-align algorithm is a term in computing advertisements, and is not a single reference to an algorithm, but a generic term for a class of methods. The method aims to find out other crowds similar to a target crowd from massive crowds based on the target crowd and realize crowd packet expansion.

elastic-search: a search engine.

bert: a natural language model.

db-scan: DBSCAN, a high density connected region based density based clustering method.

In the e-commerce field, in the process of analyzing behavior data of users giving bad comments, associated merchants and users on a shopping platform, it is found that some users try to communicate with the merchants to express and solve problems before giving bad comments, or behavior data of the users on the shopping platform can reflect that the users have dissatisfaction tendency with current commodities to a certain extent. The discovery of the dissatisfaction tendency provides possibility for the merchant to intervene in advance for the dissatisfaction users so as to improve the satisfaction degree of the users.

The specification provides a method for optimizing user shopping experience and improving merchant intervention capability to solve the problem that merchants are blinded to the reason of poor and medium comment given by users and solve the problem that poor user experience is brought by the delay of intervention after poor and medium comment given by users. The method specifically comprises the steps of firstly determining users who are over-rated and under-rated, analyzing dialogue data between the users and an object provider to determine possible discontent scenes of the users in the whole object access flow, providing a targeted solution for the object provider, predicting discontent trends and reasons of the users by analyzing the dialogue data between the users and the object provider and access behavior data of the users, and interfering the users who are possibly discontented in advance by an active service mode for the object provider, so that the object access experience of the users in the whole period is improved, and the possibility of risk of poor evaluation is reduced.

In the present specification, a data processing method is provided, and the present specification relates to a data processing apparatus, an information recommendation method, a computing device, a computer-readable storage medium, and a computer program, which are described in detail one by one in the following embodiments.

Fig. 1 is a schematic diagram illustrating a data processing procedure provided according to an embodiment of the present specification.

In the auxiliary mode of the conversation function provided in the e-commerce field at present, after the problem sent by a user is received, the problem can be searched in a knowledge base and prompted to a customer service in reception, so that the customer service can know the user requirement and the user problem can be solved quickly; through the judgment of the conversation content, the problem of the user is predicted to be met, so that the user who is met can be released by the customer service under the condition of busy, and the service efficiency of the customer service is improved.

However, most of the processing modes are applied to a conversation scene between a user and a customer service to give the customer service better service capability, so that the main solution is to improve the satisfaction degree of the user in the conversation process with the customer service, and no further attempt is made to improve the satisfaction degree of the user in the whole shopping process. Meanwhile, the processing mode mainly gives customer service prompts, improves customer service efficiency, and does not provide the capability of promoting the satisfaction degree of the user in the shopping process by pushing one key through the main contact.

Based on this, the embodiment of the present specification is mainly divided into two parts, namely discovery of an unsatisfied user and intervention of a merchant, and offline analysis of the possibility of the user dissatisfaction and configuration of a merchant solution.

1. Discontent with user discovery and merchant intervention.

Firstly, a merchant can initiate an unsatisfied user intervention process, a server side can build the crowd characteristics of private users of the merchant after receiving tasks initiated by the merchant, a series of user behaviors generated by the users on a shopping platform and dialogue data between the users and customer service are collected to build the user characteristics, the built user characteristics can be input into a user satisfaction prediction model to perform satisfaction prediction, and related users with low satisfaction distinguished by the output model are output.

Meanwhile, the server side can determine a first user giving a poor comment to the merchant and determine a second user with higher similarity to the first user so as to discover similar users and obtain a second user with low satisfaction degree and similar tendency to the first user giving the poor comment. Then, the user characteristics of the unsatisfied user, namely the second user, can be aimed at, the possible unsatisfied reason generated by the second user is judged, and the customer service is given the unsatisfied reason, or an active task is directly issued to reach the second user according to a solution configured aiming at the unsatisfied reason in advance.

2. Offline analysis of user dissatisfaction potential and merchant solution configuration.

The server side can analyze relevant conversation data of users with poor comments in real time, firstly, a part of found dissatisfaction reasons are classified according to semantic information corresponding to the conversation data, other data which cannot be classified are clustered through a text clustering algorithm, and possible dissatisfaction reasons are obtained according to clustering results.

On the basis of the possible discontent reasons, the server side can inquire a strategy for solving the discontent reasons in a question and answer knowledge base pre-configured by a merchant and send the inquiry result to the merchant, the merchant can configure solutions for the discontent reasons, can select to send related contents to a user, and can also select to send specific types of solutions to a customer service for special follow-up, and the flow of offline discontent reason discovery and merchant configuration is ended.

Fig. 2 is a flowchart illustrating a data processing method according to an embodiment of the present specification, which specifically includes the following steps:

step 202, determining an access stream generated by object access performed by a user, and acquiring to-be-processed data corresponding to different access nodes in the access stream.

Specifically, the data processing method provided in the embodiments of the present specification may be applied to a server, where the server may provide an object access service for a user through an application, that is, the user may access an object through the application, where the access manner or access behavior includes, but is not limited to, object clicking, browsing, collecting, purchasing, and the like, and the object may be a commodity or other types of tradable resources, such as virtual resources, computing resources, and the like.

The access flow, that is, data for recording a sequential relationship between access behaviors of the user for performing object access, is that, for example, the access flow generated by the user for performing commodity access may be: and accessing the commodity A, the information related to the customer service communication commodity A, the commodity B and the commodity C. In the access flow, the access commodity a, the information related to the commodity a communicated with the customer service, the access commodity B and the access commodity C are access nodes, and data related to each access node is to-be-processed data, for example, attribute information of the commodity a related to the node of the "access commodity a" and an access behavior of the user to the commodity a can be to-be-processed data corresponding to the access node; the dialogue data related to the node "information related to the customer service communication product a" may be to-be-processed data corresponding to the access node.

And 204, determining an access type corresponding to the target object access of the user according to the data to be processed.

Specifically, the target object is one of the multiple objects accessed by the user during the object access process, and different objects may belong to different object providers in the multiple objects accessed by the user, so that when the user accesses the target object provided by a certain object provider, the server can predict the access type of the user in the whole object access process according to other access behaviors involved before and after the user accesses the target object or objects provided by other accessed object providers.

The access type corresponding to the target object access by the user may be any one of several types, such as satisfaction, dissatisfaction, other or specific reasons for dissatisfaction (e.g., high price and poor quality).

Therefore, in the embodiment of the present specification, after determining an access stream generated by a user performing object access and acquiring to-be-processed data corresponding to different access nodes in the access stream, an access type corresponding to the user performing target object access may be determined according to the to-be-processed data, and the access type may be used to represent the quality of service experience or the satisfaction of an access result of the user performing target object access.

In specific implementation, the to-be-processed data includes first access data generated by the user accessing a first object, dialog data between the user and an object provider of the first object, and second access data generated by the user accessing a second object, wherein the types of the first object and the second object are the same;

correspondingly, the determining, according to the data to be processed, an access type corresponding to the target object access performed by the user includes:

constructing an access behavior sequence of the user based on the first access data, the dialogue data and the second access data;

inputting the access behavior sequence into an index value prediction model for processing to generate an index value prediction result of the target index of the user;

and determining an access type corresponding to the target object access of the user according to the index value prediction result.

Specifically, since the user can access the object through the application program of the server, the object in the application program is provided by different object providers, and the different object providers can provide the same type or similar types of objects for the user, in this case, the target object provider (any one of the object providers) is configured to improve the access satisfaction of the user in the process of performing object access on the target object provided by the user, so as to reduce user churn and increase the object access amount or transaction amount, and then the access type corresponding to the target object access performed by the user, i.e., satisfaction or dissatisfaction, can be determined according to the data to be processed generated by the user performing object access in the application program, and in the case that the access type of the user is determined to be dissatisfied, a certain remedial measure can be taken in advance to improve the user satisfaction.

In practical application, the data to be processed includes first access data generated by a user accessing a first object, dialogue data between the user and an object provider of the first object, and second access data generated by the user accessing a second object, wherein the first object and the second object are of the same type and are provided by different object providers; and then, constructing an access behavior sequence of the user based on the first access data, the dialogue data and the second access data, and specifically, sequencing the access behaviors corresponding to the access data and the dialogue data according to the sequence of the generation time of the first access data, the dialogue data and the second access data to generate the access behavior sequence. For example, the first access behavior sequence generated according to the commodity browsing record or commodity transaction record generated by the user accessing a certain shopping platform may be: the method comprises the following steps of application program home page- > commodity A detail page- > comment information of the commodity A- > commodity B home page- > commodity C home page.

After the access behavior sequence of the user is constructed, the access behavior sequence can be input into an index value prediction model for processing, an index value prediction result of a target index of the user is generated, and then an access type corresponding to target object access of the user is determined according to the index value prediction result.

Wherein, when the target index is a satisfaction index, the access type may be satisfied or unsatisfied, and based on this, according to the data to be processed, determining the access type corresponding to the target object access performed by the user includes:

inputting the data to be processed into a satisfaction prediction model for satisfaction prediction to obtain a satisfaction prediction result corresponding to target object access of the user;

and determining a target user with the satisfaction degree lower than a preset satisfaction degree threshold value according to the satisfaction degree prediction result, and determining an access type corresponding to target object access of the target user.

Specifically, under the condition that the target index is the satisfaction index, the access type corresponding to target object access of the user is determined according to the data to be processed, specifically, the data to be processed can be input into the satisfaction prediction model to perform satisfaction prediction by using the satisfaction prediction model, relevant users with low possible satisfaction distinguished by the output model are output, and the access type corresponding to target object access of the part of users is determined to be satisfactory or unsatisfactory, or the cause of the dissatisfaction is determined.

In practical application, when the object is a commodity, determining an access type corresponding to target object access of a user according to data to be processed, specifically, collecting user behavior data generated by the user on a shopping platform and dialogue data generated between the user and customer service, constructing an access behavior sequence by using the user behavior data and the dialogue data, inputting the constructed access behavior sequence into a satisfaction prediction model for satisfaction prediction, and outputting relevant users with low possible satisfaction judged by the model. When an access behavior sequence is constructed, the characteristics of a long-term behavior sequence, a short-term behavior sequence, a dialogue sequence and the like of a user can be determined based on data to be processed, the characteristics are input into a satisfaction prediction model to be encoded, a corresponding encoding vector (embedding) is obtained, then the encoding vector is processed through a multi-head attention-oriented mechanism (multi-head-orientation) of the model to extract and screen the sequences, and finally the probability that the user is not satisfied, namely an index value prediction result, is obtained through a multi-layer perception machine mlp.

The long-term behavior or the short-term behavior is actually a variable with respect to time, for example, the long-term behavior may be access data generated by a user performing object access within one month, the short-term behavior may be access data generated by a user performing object access within one day, the long-term behavior sequence is a sequence generated based on the access data of the user within one month, and the short-term behavior sequence is a sequence generated based on the access data of the user within one day. It can be seen that the essential difference between the short-term behavior and the long-term behavior is that the corresponding time lengths of the behavior data are different, and the specific time length can be determined according to the actual requirement, which is not limited herein.

In specific implementation, determining an access type corresponding to target object access performed by the user according to the data to be processed includes:

acquiring object comment information submitted by a first user aiming at a target object;

under the condition that the object comment information is determined to belong to the target type information, determining a first access type corresponding to the target object access of the first user according to the data to be processed of the first user;

calculating the similarity between the first user and the second user according to the data to be processed of the first user and the data to be processed of the second user;

and under the condition that the similarity is greater than a preset similarity threshold, determining a second access type corresponding to the target object access of the second user according to the first access type.

Further, calculating a similarity between the first user and the second user according to the data to be processed of the first user and the data to be processed of the second user includes:

determining a first object access vector of the first user according to the data to be processed of the first user;

determining a second object access vector of the second user according to the data to be processed of the second user;

and calculating the similarity between the first user and the second user according to the first object access vector and the second object access vector.

Specifically, the object comment information may be used to characterize an access type of the user to the target object, for example, if the access type includes several types, such as satisfied, dissatisfied, other or specific reasons for dissatisfaction (e.g., high price, poor quality), and the like, the access type of the user may be determined to be dissatisfied if the object comment information is negative information; or, in the case that the access type of the user is determined to be dissatisfied, further determining a specific dissatisfaction reason, and determining the dissatisfaction reason as the access type of the user; in the case where the object comment information is positive information, it can be determined that the access type of the user is satisfactory.

Therefore, in this embodiment of the present specification, after a first user accesses a target object, when an access type corresponding to target object access by a second user needs to be determined, when object comment information is submitted for the target object by the first user, the object comment information submitted for the target object by the first user may be obtained first, then, when it is determined that the object comment information belongs to target type information, a first access type corresponding to the target object access by the first user is determined according to data to be processed generated in a process in which the first user accesses the target object, then, a second user with a higher similarity to the first user may be determined, and a second access type for the target object access by the second user is determined according to the first access type of the first user, and preferably, when it is determined that a similarity between the first user and the second user is greater than a preset similarity threshold, the first access type of the first user may be determined as the second access type for the target object access by the second user.

When the similarity between the first user and the second user is calculated according to the data to be processed of the first user and the data to be processed of the second user, the object access vectors of the users are determined according to the data to be processed, and the similarity between different users is determined by calculating the distance between the object access vectors.

In practical application, if the target object is a commodity, the first user may perform evaluation on the commodity after purchasing the commodity, that is, submit object comment information, where the evaluation is usually classified as good, medium or poor, and the target type information is medium or poor, so that, when the evaluation submitted by the first user is determined as medium or poor, it may be determined that the access type of the first user to the commodity is dissatisfied, and in this case, it may further determine, according to the commodity access data of the first user, a dissatisfaction scenario, that is, a dissatisfaction reason, of the first user, and then determine a second user with a higher similarity to the first user, where the second user is a user who is accessing the commodity but has not made a purchase order, and determine the dissatisfaction reason of the first user as an access type corresponding to a commodity transaction of the second user, so as to obtain a low-degree user having a similar tendency to the user who has given medium or poor.

In the embodiment of the specification, a first access type of a first user for performing target object access is determined according to-be-processed data of the first user who has given object comment information, a second user with higher similarity to the first user is determined, a second access type corresponding to the target object access of the second user is determined according to the first access type of the first user, and based on the second access type, a second user who may also give object comment information the same as or similar to the object comment information of the first user is determined in a similar user discovery mode based on the first user, so that the second user is intervened in advance, and therefore service experience or object access satisfaction of the second user in the whole object access process is improved.

In addition, a graph structure can be constructed according to the data to be processed of the first user, wherein the graph structure comprises a user node corresponding to the first user and a data node corresponding to the data to be processed of the first user;

correspondingly, the calculating the similarity between the first user and the second user according to the data to be processed of the first user and the data to be processed of the second user includes:

updating the graph structure according to the data to be processed of the second user to generate a target graph structure;

obtaining a first vector quantization representation of a user node corresponding to the first user and a second vector quantization representation of a user node corresponding to the second user in the target graph structure based on a preset network embedded learning model;

and calculating the similarity between the first user and the second user according to the first vector quantization representation and the second vector quantization representation.

Further, obtaining a first vector quantization representation of a user node corresponding to the first user in the target graph structure based on a preset network-embedded learning model includes:

adopting a random walk algorithm to perform sequence sampling on user nodes corresponding to the first user in the target graph structure to generate a first access sequence;

vectorizing the first access sequence based on a preset network embedded learning model.

In particular, graph (Graph) is a more complex data structure than linear tables and trees. And the graph structure is a many-to-many relationship between study data elements. In this structure, there may be a relationship between any two elements, that is, the relationship between nodes may be arbitrary, and any element in the graph may be related; look-like, similar population expansion, is based on seed users, through certain algorithm assessment model, finds more similar population's technology that possess potential relevance. It is worth noting that look-align is not a specific algorithm, but a generic name of a class of methods, and the class of methods comprehensively uses a plurality of technologies, such as collaborative filtering, node2vec and the like, to finally achieve the purpose of user expansion.

In the embodiment of the present specification, a user similar to the determined population with low satisfaction degree may be determined through a look-like algorithm and the like, and is used as a user with low satisfaction degree obtained through expansion, wherein a graph structure may be constructed according to the data to be processed of the first user, specifically, a user identifier of the first user is used as a first node, a data identifier of each data to be processed is used as a second node, and an association relationship between the first user and the data to be processed is used as an edge between the first node and the second node to construct the graph structure.

After incremental data exist, namely newly generated data to be processed of a second user exist, the graph structure can be updated according to the data to be processed of the second user, and a target graph structure is generated; then, a random walk algorithm can be adopted to carry out sequence sampling on user nodes corresponding to a first user in a target graph structure to generate a first access sequence, and vectorization representation is carried out on the first access sequence based on a Node2vec network embedded learning model; or vectorizing the first access sequence based on a Deep Walk network embedded learning model, and calculating the similarity between the first user and the second user according to a first vectorized representation corresponding to the first user and a second vectorized representation corresponding to the second user. The similarity between the first user and the second user is calculated according to the first vector quantization representation and the second vector quantization representation, specifically, the Euclidean distance or cosine similarity between the first vector quantization representation and the second vector quantization representation can be calculated, so that the similarity between the first user and the second user is determined according to the size of the calculation result, and in practical application, the smaller the calculation result is, the larger the similarity between the first user and the second user is.

In practical application, the first access sequence is vectorized and expressed based on a Node2vec network embedded learning model, that is, the first access sequence is vectorized and expressed based on a Word2vec SkipGram framework through the Node2vec network embedded learning model.

In one or more embodiments of the present specification, based on a Node2vec network embedded learning model, a random walk algorithm (random walk) is adopted to convert a user Node in a constructed target graph structure into a sampled access sequence, and then based on a SkipGram frame in a Word2vec model, probabilistic learning and inference are performed on the sampled access sequence, so as to finally obtain a vectorized representation of the user Node in the target graph structure. The node vectorization expression obtained through network embedding learning can enrich the relation among nodes and improve the speed and the effect of classification.

A schematic diagram of a data processing process provided in an embodiment of the present specification is shown in fig. 3a, a satisfaction task may be created by a merchant, and a process of unsatisfied user intervention is initiated, after receiving the task initiated by the merchant, a server side may construct a crowd characteristic for a private user of the merchant, collect a series of user behaviors generated by a user on a shopping platform and dialogue data between the user and customer service, so as to construct a user characteristic, where the constructed user characteristic is input to a user satisfaction prediction model to perform satisfaction prediction, and output a relevant user with low satisfaction distinguished by the model.

Meanwhile, the server side can determine a first user giving a poor comment to the merchant and determine a second user with higher similarity to the first user so as to discover similar users and obtain a second user with low satisfaction degree and similar tendency to the first user giving the poor comment. Then, the user characteristics of the unsatisfied user, namely the second user, can be aimed at, the possible unsatisfied reason generated by the second user is judged, and the customer service is given the unsatisfied reason, or an active task is directly issued to reach the user according to a solution configured aiming at the unsatisfied reason in advance.

In specific implementation, the data to be processed includes object comment information submitted by the user for a target object, and includes dialogue data between the user and an object provider of the target object;

and under the condition that the object comment information is determined to belong to the target type information, determining an access type corresponding to the target object access of the user according to the dialogue data.

Specifically, as described above, after the user accesses the target object, the object comment information may be submitted for the target object, where the object comment information may be used to characterize the access type of the user for the target object, for example, if the access type includes several types, such as satisfied types, dissatisfied types, other or specific dissatisfied reasons (e.g., high price and poor quality), if the access type is negative information, it may be determined that the access type of the user is dissatisfied; or, in the case that the access type of the user is determined to be dissatisfied, further determining a specific dissatisfaction reason, and determining the dissatisfaction reason as the access type of the user; in the case where the object comment information is positive information, it can be determined that the access type of the user is satisfactory.

Therefore, in the embodiment of the present specification, after the user accesses the target object, when the object comment information is submitted for the target object, and when an access type corresponding to the target object access by the user needs to be determined, the object comment information submitted for the target object by the user may be obtained first, and then, when it is determined that the object comment information belongs to the object type information, the access type corresponding to the target object access by the user is determined according to the to-be-processed data generated between the user and the object provider of the target object during the process of accessing the target object by the user.

Based on the method, under the condition that the user submits the object comment information aiming at the target object and the object comment information is determined to belong to the target type information, the access type corresponding to the target object access of the user can be further determined according to the dialogue data generated between the user and the object provider, so that the information recommendation can be performed on the user according to the access type, and therefore the service experience or the object access satisfaction of the user in the whole object access process can be improved.

In addition, before determining the access type corresponding to the target object access performed by the user, offline analysis may be performed on the access type that may exist in the user, and solution configuration may be performed, which may specifically be implemented by the following method:

acquiring object comment information submitted by at least two first users aiming at a target object;

under the condition that the object comment information is determined to belong to the target type information, acquiring dialogue data between the at least two first users and an object provider of the target object;

determining access types corresponding to target object access of the at least two first users according to semantic information corresponding to the dialogue data; alternatively, the first and second liquid crystal display panels may be,

and clustering the dialogue data, and determining the access types corresponding to the target object access of the at least two first users according to the clustering result.

Specifically, since the user may attempt to communicate with the object provider to express and solve the problem before submitting the object comment information, the dialogue data between the user and the object provider may reflect, to some extent, the user's dissatisfaction tendency with the target object. The discovery of the dissatisfaction tendency provides possibility for the object provider to recommend information to the dissatisfaction users so as to improve the satisfaction degree of the users.

Based on this, in the embodiment of the present specification, before determining the access type corresponding to the target object access performed by the second user, different access types that may exist are determined according to the object comment information of the first user, and then the access type corresponding to the second user is determined according to the similarity between the first user and the second user.

Specifically, under the condition that the object comment information of the first user is determined to belong to the object type information, the dialogue data between the first user and the object provider of the object can be acquired, and the access type corresponding to the object access of the first user is determined according to the semantic information of the dialogue data; for example, if the target object is a commodity and the dialogue data is "why it has not been received for that long time", it may be determined that the access type thereof is slow in logistics; if the conversation data is 'the packaging box is broken when received', the access type of the conversation data can be determined to be packaging damage; however, if the session data is "can help me recommend", the access type cannot be determined.

In practical application, the dialogue data can be input into a language model such as bert and the like to be classified and predicted, and whether the dialogue data belongs to the determined access type or not can be predicted; for the dialogue data of which the access type cannot be determined, clustering is performed on the dialogue data through a clustering algorithm such as dbscan to obtain corresponding clustering clusters, and each clustering cluster can represent one access type.

Another schematic diagram of a data processing process provided in this specification is shown in fig. 3b, where the server may analyze, in real time, relevant session data of a user who has generated bad comments, obtain a part of classifications of found dissatisfaction causes through semantic information corresponding to the session data, and cluster other data that cannot be classified through a text clustering algorithm, and obtain possible dissatisfaction causes according to a clustering result.

On the basis of obtaining the possible discontent reasons, the service end can inquire a strategy for solving the discontent reasons in a question and answer knowledge base configured by a merchant in advance through an elastic-search frame and the like, and sends an inquiry result to the merchant, the merchant can configure solutions for the discontent reasons, can select to send related contents to a user, and can also select to send specific types of solutions to a customer service for special follow-up, and the flow of offline discontent reason discovery and merchant configuration is ended.

And step 206, determining information to be recommended corresponding to the target access type under the condition that the access type belongs to the target access type.

Specifically, if the access type includes any one of several types such as satisfaction, dissatisfaction, other reasons for dissatisfaction, or specific reasons for dissatisfaction (e.g., high price and poor quality), the target access type is the dissatisfaction or specific dissatisfaction reason, and if it is determined that the access type of the user performing the target object access belongs to the target access type, information recommendation may be performed for the user in order to improve the user satisfaction, for example, if it is determined that the dissatisfaction reason of the user is high price, a coupon may be issued for the user, or if the dissatisfaction reason of the user is poor quality, services such as a gift delivery, discount, or refund may be provided for the user.

And step 208, displaying the information to be recommended to the user through an information operation interface.

Specifically, after the information to be recommended is determined, the information to be recommended can be displayed to the user through an information operation interface of the server, or the information to be recommended can be sent to the user through other online (for example, a customer service chat window) or offline (for example, express delivery, telephone call, short message) modes, so that the user is informed of the information to be recommended in time, and if the user accepts a processing strategy in the information to be recommended, for example, a coupon is used, the satisfaction degree in the user object access process can be effectively improved.

The embodiment of the specification analyzes the satisfaction degree of the user in the whole shopping process from an off-line analysis stage, and simultaneously analyzes the conversation data and the access behavior data of the whole process, not only the current conversation scene, when the dissatisfaction prediction of the user is carried out, so that the possibility of bringing the whole-period shopping experience to the user is ensured. In addition, the embodiments of the present specification can configure a solution that can be used for the merchant in the offline configuration stage, which not only improves the customer service reception experience perceived by the user during the conversation process, but also can give better and more thorough service to the user in the situation that the user may generate dissatisfaction by means of the capability of the merchant for active service.

It can be seen that, in the embodiment of the present specification, the dialogue data of the buyer with poor comment and medium comment is analyzed, the possible discontent scene of the buyer in the whole shopping process is combed out, the capability of providing targeted solution configuration for the merchant is presented and provided by the merchant, in addition, the discontent tendency and the reason of the buyer can be predicted by analyzing the dialogue and behavior of the buyer, the merchant can intervene in advance on the buyer which may be discontented through active service, the shopping experience of the user in the whole period is improved, and the possibility of the risk of poor comment is reduced.

In the process of object access by a user, the embodiment of the specification predicts the access type of the user by analyzing the dialogue data and/or access behavior data (to-be-processed data) corresponding to different access nodes in the generated access stream, so that an object provider can intervene in advance on the user according to the access type, namely, information to be recommended is sent to the user, and the service experience or object access satisfaction of the user in the whole object access process is improved through the information to be recommended.

The following description will further explain the data processing method provided in this specification by taking an application of the data processing method in a commodity transaction scenario as an example with reference to fig. 4. Fig. 4 shows a flowchart of a processing procedure of a data processing method according to an embodiment of the present specification, which specifically includes the following steps. A

Step 402, obtaining a commodity evaluation submitted by a first user for a target commodity.

And step 404, under the condition that the commodity evaluation belongs to the medium and poor evaluation, determining conversation data between the first user and a merchant to which the target commodity belongs.

Step 406, determining a first access type corresponding to the target commodity access of the first user according to the dialogue data.

And 408, determining an access flow generated by commodity access of the first user, and acquiring to-be-processed data corresponding to different access nodes in the access flow.

Step 410, determining an access flow generated by the second user for commodity access, and acquiring to-be-processed data corresponding to different access nodes in the access flow.

In step 412, the similarity between the first user and the second user is calculated according to the data to be processed of the first user and the data to be processed of the second user.

And step 414, determining a second access type corresponding to the target commodity access of the second user according to the first access type under the condition that the similarity is greater than a preset similarity threshold.

And step 416, determining information to be recommended corresponding to the second access type.

And 418, displaying the information to be recommended to the second user through the conversation interaction interface.

In the commodity access process of the user, the access type of the user is predicted by analyzing the conversation data and/or the access behavior data (to-be-processed data) corresponding to different access nodes in the generated access stream, so that a merchant can intervene the user in advance according to the access type, namely, the to-be-recommended information is sent to the user, and the service experience or satisfaction of the user in the whole commodity access process is improved through the to-be-recommended information.

Corresponding to the above method embodiment, the present specification further provides an embodiment of a data processing apparatus, and fig. 5 shows a schematic structural diagram of a data processing apparatus provided in an embodiment of the present specification. As shown in fig. 5, the apparatus includes:

an obtaining module 502, configured to determine an access stream generated by a user performing object access, and obtain to-be-processed data corresponding to different access nodes in the access stream;

a first determining module 504, configured to determine, according to the to-be-processed data, an access type corresponding to target object access performed by the user;

a second determining module 506, configured to determine information to be recommended corresponding to the target access type if the access type belongs to the target access type;

a presentation module 508 configured to present the information to be recommended to the user through an information operation interface.

Optionally, the data to be processed includes first access data generated by the user accessing a first object, dialog data between the user and an object provider of the first object, and second access data generated by the user accessing a second object, where the first object and the second object are of the same type;

accordingly, the first determining module 504 is further configured to:

Optionally, the first determining module 504 is further configured to:

under the condition that the object comment information is determined to belong to the target type information, determining a first access type corresponding to target object access of the first user according to the data to be processed of the first user;

Optionally, the first determining module 504 is further configured to:

Optionally, the data processing apparatus further includes a construction module configured to:

constructing a graph structure according to the data to be processed of the first user, wherein the graph structure comprises a user node corresponding to the first user and a data node corresponding to the data to be processed of the first user;

accordingly, the first determining module 504 is further configured to:

Optionally, the first determining module 504 is further configured to:

Optionally, the data to be processed includes object comment information submitted by the user for a target object, and includes dialogue data between the user and an object provider of the target object;

accordingly, the first determining module 504 is further configured to:

Optionally, the data processing apparatus further comprises a clustering module configured to:

determining access types corresponding to target object access of the at least two first users according to semantic information corresponding to the dialogue data; alternatively, the first and second electrodes may be,

Optionally, the first determining module 504 is further configured to:

and determining a target user with the satisfaction degree lower than a preset satisfaction degree threshold according to the satisfaction degree prediction result, and determining an access type corresponding to the target user for accessing the target object.

The above is a schematic configuration of a data processing apparatus of the present embodiment. It should be noted that the technical solution of the data processing apparatus belongs to the same concept as the technical solution of the data processing method, and for details that are not described in detail in the technical solution of the data processing apparatus, reference may be made to the description of the technical solution of the data processing method.

Fig. 6 is a flowchart illustrating an information recommendation method according to an embodiment of the present specification, which specifically includes the following steps:

step 602, determining an access flow generated by a user for commodity access, and acquiring to-be-processed data corresponding to different access nodes in the access flow.

Step 604, determining an access type corresponding to the target commodity access of the user according to the data to be processed.

Step 606, determining information to be recommended corresponding to the target access type under the condition that the access type belongs to the target access type.

And 608, displaying the information to be recommended to the user through a conversation interactive interface.

In one embodiment of the description, an access flow generated by a user for commodity access is determined, to-be-processed data corresponding to different access nodes in the access flow is acquired, an access type corresponding to target commodity access of the user is determined according to the to-be-processed data, to-be-recommended information corresponding to the target access type is determined under the condition that the access type belongs to the target access type, and the to-be-recommended information is displayed to the user through a conversation interaction interface.

The foregoing is a schematic scheme of an information recommendation method according to this embodiment. It should be noted that the technical solution of the information recommendation method and the technical solution of the data processing method belong to the same concept, and details that are not described in detail in the technical solution of the information recommendation method can be referred to the description of the technical solution of the data processing method.

Corresponding to the above method embodiment, the present specification further provides an information recommendation apparatus embodiment, and fig. 7 shows a schematic structural diagram of an information recommendation apparatus provided in an embodiment of the present specification. As shown in fig. 7, the apparatus includes:

an obtaining module 702 is configured to determine an access flow generated by a user performing commodity access, and obtain to-be-processed data corresponding to different access nodes in the access flow.

A first determining module 704 configured to determine, according to the to-be-processed data, an access type corresponding to target commodity access performed by the user.

A second determining module 706 configured to determine information to be recommended corresponding to the target access type if the access type belongs to the target access type.

A presentation module 708 configured to present the information to be recommended to the user through a dialog interactive interface.

The above is a schematic scheme of an information recommendation apparatus of the present embodiment. It should be noted that the technical solution of the information recommendation apparatus and the technical solution of the information recommendation method described above belong to the same concept, and for details that are not described in detail in the technical solution of the information recommendation apparatus, reference may be made to the description of the technical solution of the information recommendation method described above.

FIG. 8 illustrates a block diagram of a computing device 800, according to one embodiment of the present description. The components of the computing device 800 include, but are not limited to, memory 810 and a processor 820. The processor 820 is coupled to the memory 810 via a bus 830, and the database 850 is used to store data.

Computing device 800 also includes access device 840, access device 840 enabling computing device 800 to communicate via one or more networks 860. Examples of such networks include the Public Switched Telephone Network (PSTN), a Local Area Network (LAN), a Wide Area Network (WAN), a Personal Area Network (PAN), or a combination of communication networks such as the internet. Access device 840 may include one or more of any type of network interface (e.g., a Network Interface Card (NIC)) whether wired or wireless, such as an IEEE802.11 Wireless Local Area Network (WLAN) wireless interface, a worldwide interoperability for microwave access (Wi-MAX) interface, an ethernet interface, a Universal Serial Bus (USB) interface, a cellular network interface, a bluetooth interface, a Near Field Communication (NFC) interface, and so forth.

In one embodiment of the present description, the above-described components of computing device 800, as well as other components not shown in FIG. 8, may also be connected to each other, such as by a bus. It should be understood that the block diagram of the computing device architecture shown in FIG. 8 is for purposes of example only and is not limiting as to the scope of the description. Those skilled in the art may add or replace other components as desired.

Computing device 800 may be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., tablet, personal digital assistant, laptop, notebook, netbook, etc.), a mobile phone (e.g., smartphone), a wearable computing device (e.g., smartwatch, smartglasses, etc.), or other type of mobile device, or a stationary computing device such as a desktop computer or PC. Computing device 800 may also be a mobile or stationary server.

The processor 820 is configured to execute computer-executable instructions, and when executed by the processor, the computer-executable instructions implement the steps of the data processing method or the information recommendation method.

The foregoing is a schematic diagram of a computing device of the present embodiment. It should be noted that the technical solution of the computing device and the technical solution of the data processing method or the information recommendation method described above belong to the same concept, and details that are not described in detail in the technical solution of the computing device can be referred to the description of the technical solution of the data processing method or the information recommendation method described above.

An embodiment of the present specification further provides a computer-readable storage medium, which stores computer-executable instructions, and when the computer-executable instructions are executed by a processor, the steps of the data processing method or the information recommendation method are implemented.

The above is an illustrative scheme of a computer-readable storage medium of the present embodiment. It should be noted that the technical solution of the storage medium belongs to the same concept as the technical solution of the data processing method or the information recommendation method, and details that are not described in detail in the technical solution of the storage medium can be referred to the description of the technical solution of the data processing method or the information recommendation method.

An embodiment of the present specification further provides a computer program, wherein when the computer program is executed in a computer, the computer program causes the computer to execute the steps of the data processing method or the information recommendation method.

The above is an illustrative scheme of a computer program of the present embodiment. It should be noted that the technical solution of the computer program is the same as the technical solution of the data processing method or the information recommendation method, and details that are not described in detail in the technical solution of the computer program can be referred to the description of the technical solution of the data processing method or the information recommendation method.

The foregoing description has been directed to specific embodiments of this disclosure. Other embodiments are within the scope of the following claims. In some cases, the actions or steps recited in the claims may be performed in a different order than in the embodiments and still achieve desirable results. In addition, the processes depicted in the accompanying figures do not necessarily require the particular order shown, or sequential order, to achieve desirable results. In some embodiments, multitasking and parallel processing may also be possible or may be advantageous. The computer instructions comprise computer program code which may be in the form of source code, object code, an executable file or some intermediate form, or the like. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, usb disk, removable hard disk, magnetic disk, optical disk, computer Memory, read-Only Memory (ROM), random Access Memory (RAM), electrical carrier wave signals, telecommunications signals, software distribution medium, and the like. It should be noted that the computer readable medium may contain content that is subject to appropriate increase or decrease as required by legislation and patent practice in jurisdictions, for example, in some jurisdictions, computer readable media does not include electrical carrier signals and telecommunications signals as is required by legislation and patent practice.

It should be noted that, for the sake of simplicity, the foregoing method embodiments are described as a series of combinations of acts, but it should be understood by those skilled in the art that the embodiments are not limited by the described order of acts, as some steps may be performed in other orders or simultaneously according to the embodiments. Further, those skilled in the art should also appreciate that the embodiments described in this specification are preferred embodiments and that acts and modules referred to are not necessarily required for an embodiment of the specification.

In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and for parts that are not described in detail in a certain embodiment, reference may be made to related descriptions of other embodiments.

The preferred embodiments of the present specification disclosed above are intended only to aid in the description of the specification. Alternative embodiments are not exhaustive and do not limit the invention to the precise embodiments described. Obviously, many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described in order to best explain the principles of the embodiments and the practical application, and to thereby enable others skilled in the art to best understand the specification and utilize the specification. The specification is limited only by the claims and their full scope and equivalents.

Claims

1. A method of data processing, comprising:

2. The data processing method according to claim 1, wherein the data to be processed includes first access data generated by the user accessing a first object, dialog data between the user and an object provider of the first object, and second access data generated by the user accessing a second object, wherein the first object and the second object are of the same type;

3. The data processing method according to claim 1, wherein the determining, according to the data to be processed, an access type corresponding to target object access performed by the user comprises:

4. The data processing method according to claim 3, wherein the calculating the similarity between the first user and the second user according to the data to be processed of the first user and the data to be processed of the second user comprises:

5. The data processing method of claim 3, further comprising:

6. The data processing method according to claim 5, wherein the obtaining a first vector quantization representation of a user node corresponding to the first user in the target graph structure based on a preset network-embedded learning model comprises:

sequence sampling is carried out on user nodes corresponding to the first user in the target graph structure by adopting a random walk algorithm, and a first access sequence is generated;

7. The data processing method according to claim 1, wherein the data to be processed includes object comment information submitted by the user for a target object, and includes dialogue data between the user and an object provider of the target object;

8. The data processing method of claim 1, further comprising:

determining access types corresponding to the target object access of the at least two first users according to semantic information corresponding to the dialogue data; alternatively, the first and second electrodes may be,

9. The data processing method according to claim 1, wherein the determining, according to the data to be processed, an access type corresponding to target object access performed by the user comprises:

10. A data processing apparatus comprising:

the acquisition module is configured to determine an access stream generated by object access of a user and acquire to-be-processed data corresponding to different access nodes in the access stream;

11. An information recommendation method, comprising:

12. A computing device, comprising:

a memory and a processor;

the memory is configured to store computer-executable instructions and the processor is configured to execute the computer-executable instructions, which when executed by the processor implement the steps of the data processing method of any one of claims 1 to 9 or the information recommendation method of claim 11.

13. A computer-readable storage medium storing computer-executable instructions which, when executed by a processor, implement the steps of the data processing method of any one of claims 1 to 9 or the information recommendation method of claim 11.