Artificial intelligence-based information pushing method and device, equipment and storage medium

Abstract

The application belongs to the field of artificial intelligence and the field of financial technology, and relates to an information pushing method based on artificial intelligence, which comprises the following steps: receiving an information query request corresponding to a target customer and triggered by a target operator in a work process; obtaining work characteristic information corresponding to the work process; generating a prediction probability value of the work characteristic information corresponding to various classification label information through an information prediction model; determining target classification label information meeting the target classification label information from the classification label information based on the prediction probability value; obtaining target query information corresponding to the target classification label information; and pushing the target query information to the target operator. The application also provides an information pushing device based on artificial intelligence, a computer device and a storage medium. In addition, the application also relates to blockchain technology, and the target query information can be stored in the blockchain. The application can be applied to an information query scene in the financial field, and can effectively improve the processing efficiency of the target operator in obtaining important customer information.

Description

Technical Field

[0001] This application relates to the fields of artificial intelligence development technology and financial technology, and in particular to information push methods, devices, computer equipment and storage media based on artificial intelligence. Background Technology

[0002] Financial insurance companies typically have customer service representatives to facilitate business interactions with clients. Currently, during customer service operations, such as those handling auto insurance or non-auto insurance transactions, all customer information is categorized according to system functionality rather than being designed specifically for the workflow. This results in the same complete customer information being displayed at different stages of customer contact.

[0003] Because agent operations involve numerous steps, and the customer information required for each step varies, agents must manually click through all customer information displayed on a pre-assembled single page to find the relevant information for the current task. However, due to the large volume of customer information, often involving hundreds of fields, the current method of manually clicking through customer information is time-consuming, inefficient, and consequently leads to low agent productivity. Summary of the Invention

[0004] The purpose of this application is to propose an information push method, device, computer equipment, and storage medium based on artificial intelligence, in order to solve the technical problem that the existing method of manually clicking on customer information to retrieve information by agents takes a lot of time, has low information retrieval processing efficiency, and thus leads to low work efficiency of agents.

[0005] To address the aforementioned technical problems, this application provides an information push method based on artificial intelligence, employing the following technical solution:

[0006] Determine whether an information query request corresponding to the target customer has been received by the target agent during the operation process;

[0007] If so, based on the preset information type, obtain the operation feature information corresponding to the operation process;

[0008] The task feature information is input into a pre-trained information prediction model, and the information prediction model is used to predict the task feature information to generate prediction probability values ​​of the task feature information corresponding to various classification label information.

[0009] Based on the predicted probability values, target classification label information that meets the preset importance conditions is determined from all the classification label information;

[0010] Obtain the target query information of the target customer corresponding to the target category tag information;

[0011] The target query information is pushed to the target agent's work interface.

[0012] Furthermore, the step of determining the target classification label information that meets the preset importance condition from all the classification label information based on each of the predicted probability values ​​specifically includes:

[0013] Obtain the preset probability threshold;

[0014] Each of the predicted probability values ​​is compared with the probability threshold, and a specified probability threshold greater than the probability threshold is selected from all the probability thresholds.

[0015] Filter out the specified category label information that corresponds to the specified probability threshold from all the category label information;

[0016] The specified category label information is used as the target category label information.

[0017] Furthermore, the step of pushing the target query information to the target agent's work interface specifically includes:

[0018] Obtain the target probability value corresponding to each of the target classification label information;

[0019] Based on the target probability value, generate the information priority of each target query information;

[0020] Get the preset display page;

[0021] Based on the information priority of each of the target query information, the display position of each of the target query information is determined on the display page;

[0022] Based on the display position of each of the target query information, each of the target query information is filled into the display page accordingly to obtain the target display page;

[0023] The target display page is pushed to the target agent's work interface.

[0024] Furthermore, before the step of inputting the job feature information into a pre-trained information prediction model, and using the information prediction model to predict the job feature information to generate predicted probability values ​​for the job feature information corresponding to various classification label information, the method further includes:

[0025] Obtain pre-collected historical operation feature information;

[0026] The historical operation feature information is preprocessed to obtain corresponding sample data;

[0027] The sample data is divided into a training set and a test set;

[0028] Multiple initial models were constructed based on the XGBoost algorithm.

[0029] The training set is used to train each of the first initial models to obtain multiple trained second initial models.

[0030] Each of the second initial models is used to perform prediction processing on the test set to obtain the prediction results output by each of the second initial models;

[0031] Based on the prediction results output by each of the second initial models, model screening is performed on each of the second initial models to determine the information prediction model from all the second initial models.

[0032] Furthermore, the step of preprocessing the historical operation feature information to obtain the corresponding sample data specifically includes:

[0033] The historical job feature information is cleaned to obtain the corresponding first job feature information;

[0034] The first job feature information is converted to obtain the corresponding second job feature information;

[0035] The second job feature information is used as the sample data.

[0036] Furthermore, the step of screening each second initial model based on the prediction results output by each second initial model, and determining the information prediction model from all second initial models, specifically includes:

[0037] Based on the prediction results output by each of the second initial models, the second initial models are evaluated to obtain the processing effect value of each of the second initial models;

[0038] Numerical analysis was performed on all the processing effect values, and the third initial model with the highest processing effect value was selected from all the second initial models.

[0039] The third initial model is used as the information prediction model.

[0040] Furthermore, after the step of performing model screening on each of the second initial models based on the prediction results output by each of the second initial models, and determining the information prediction model from all the second initial models, the method further includes:

[0041] The information prediction model is then processed for publication.

[0042] After completing the publishing process of the information prediction model, the information prediction model is stored.

[0043] Obtain a fourth initial model from all of the second initial models, excluding the information prediction model;

[0044] The fourth initial model is then deleted.

[0045] To address the aforementioned technical problems, this application also provides an information push device based on artificial intelligence, employing the following technical solution:

[0046] The judgment module is used to determine whether an information query request corresponding to the target customer has been received by the target agent during the operation.

[0047] The first acquisition module is used to acquire, if so, job feature information corresponding to the job process based on a preset information type;

[0048] The prediction module is used to input the job feature information into a pre-trained information prediction model, and to perform prediction processing on the job feature information through the information prediction model to generate prediction probability values ​​of the job feature information corresponding to various classification label information.

[0049] The first determining module is used to determine, based on each of the predicted probability values, target classification label information that meets the preset importance conditions from all the classification label information;

[0050] The second acquisition module is used to acquire the target query information of the target customer corresponding to the target category label information;

[0051] The push module is used to push the target query information to the target agent's work interface.

[0052] To address the aforementioned technical problems, this application also provides a computer device that employs the following technical solution:

[0053] Determine whether an information query request corresponding to the target customer has been received by the target agent during the operation process;

[0054] If so, based on the preset information type, obtain the operation feature information corresponding to the operation process;

[0055] The task feature information is input into a pre-trained information prediction model, and the information prediction model is used to predict the task feature information to generate prediction probability values ​​of the task feature information corresponding to various classification label information.

[0056] Based on the predicted probability values, target classification label information that meets the preset importance conditions is determined from all the classification label information;

[0057] Obtain the target query information of the target customer corresponding to the target category tag information;

[0058] The target query information is pushed to the target agent's work interface.

[0059] To address the aforementioned technical problems, this application also provides a computer-readable storage medium, employing the technical solution described below:

[0060] Determine whether an information query request corresponding to the target customer has been received by the target agent during the operation process;

[0061] If so, based on the preset information type, obtain the job feature information corresponding to the job process;

[0062] The task feature information is input into a pre-trained information prediction model, and the information prediction model is used to predict the task feature information to generate prediction probability values ​​of the task feature information corresponding to various classification label information.

[0063] Based on the predicted probability values, target classification label information that meets the preset importance conditions is determined from all the classification label information;

[0064] Obtain the target query information of the target customer corresponding to the target category tag information;

[0065] The target query information is pushed to the target agent's work interface.

[0066] Compared with the prior art, the embodiments of this application have the following main advantages:

[0067] This application embodiment first determines whether an information query request corresponding to a target customer, triggered by a target agent during the operation, has been received. If so, based on a preset information type, operation feature information corresponding to the operation is obtained. Then, the operation feature information is input into a pre-trained information prediction model, and the operation feature information is predicted by the information prediction model to generate predicted probability values ​​for the operation feature information corresponding to various classification tag information. Subsequently, based on each predicted probability value, target classification tag information that meets preset importance conditions is determined from all the classification tag information. Further, the target query information of the target customer corresponding to the target classification tag information is obtained. Finally, the target query information is pushed to the target agent's work interface. In this embodiment, when a target agent receives an information query request corresponding to a target customer triggered during a task, the system automatically acquires task feature information corresponding to the task. Then, based on a pre-trained information prediction model, the system performs predictive processing on the task feature information, generating predicted probability values ​​for various category tag information corresponding to the task feature information. Based on these predicted probability values, the system determines the target category tag information that meets the importance criteria, thereby acquiring the target customer's query information corresponding to the target category tag information. Finally, the target query information is pushed to the target agent's work interface to satisfy the target agent's information access needs. The use of the information prediction model accurately predicts the target query information corresponding to the target customer, ensuring the accuracy of the generated query information. It also effectively reduces the time spent by the target agent searching for query information, improving the processing efficiency of the target agent in obtaining effective customer information, thus enhancing the target agent's work efficiency and user experience. Attached Figure Description

[0068] To more clearly illustrate the solutions in this application, the accompanying drawings used in the description of the embodiments of this application will be briefly introduced below. Obviously, the accompanying drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0069] Figure 1 This is an exemplary system architecture diagram to which this application can be applied;

[0070] Figure 2 A flowchart of an embodiment of the AI-based information push method according to this application;

[0071] Figure 3 This is a schematic diagram of a structure of an embodiment of an AI-based information push device according to this application;

[0072] Figure 4 This is a schematic diagram of the structure of one embodiment of the computer device according to this application. Detailed Implementation

[0073] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains; the terminology used herein in the specification of the application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application; the terms "comprising" and "having," and any variations thereof, in the specification, claims, and foregoing drawings of this application, are intended to cover non-exclusive inclusion. The terms "first," "second," etc., in the specification, claims, or foregoing drawings of this application are used to distinguish different objects, not to describe a particular order.

[0074] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0075] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0076] like Figure 1 As shown, system architecture 100 may include terminal devices 101, 102, and 103, a network 104, and a server 105. Network 104 serves as the medium for providing communication links between terminal devices 101, 102, and 103 and server 105. Network 104 may include various connection types, such as wired or wireless communication links, or fiber optic cables, etc.

[0077] Users can use terminal devices 101, 102, and 103 to interact with server 105 via network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, and 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social media platform software, etc.

[0078] Terminal devices 101, 102, and 103 can be various electronic devices with displays and support web browsing, including but not limited to smartphones, tablets, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 players (Moving Picture Experts Group Audio Layer IV), laptops, and desktop computers, etc.

[0079] Server 105 can be a server that provides various services, such as a backend server that supports the pages displayed on terminal devices 101, 102, and 103.

[0080] It should be noted that the information push method based on artificial intelligence provided in this application embodiment is generally executed by a server / terminal device, and correspondingly, the information push device based on artificial intelligence is generally set in the server / terminal device.

[0081] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0082] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0083] It should be understood that Figure 1 The number of terminal devices, networks, and servers shown is merely illustrative. Depending on implementation needs, any number of terminal devices, networks, and servers can be included.

[0084] Continue to refer to Figure 2The flowchart illustrates an embodiment of the AI-based information push method according to this application. The order of steps in the flowchart can be changed, and some steps can be omitted, depending on different needs. The AI-based information push method provided in this application embodiment can be applied to any scenario requiring customer information inquiry, and thus can be applied to products in these scenarios, such as customer information inquiry during agent operations in the financial insurance field. The AI-based information push method includes the following steps:

[0085] Step S201: Determine whether an information query request corresponding to the target customer has been received by the target agent during the operation.

[0086] In this embodiment, the information push method based on artificial intelligence runs on an electronic device (e.g., Figure 1 The server / terminal device shown can obtain information query requests via wired or wireless connections. It should be noted that the aforementioned wireless connection methods may include, but are not limited to, 3G / 4G / 5G connections, WiFi connections, Bluetooth connections, WiMAX connections, Zigbee connections, UWB (ultra-wideband) connections, and other currently known or future-developed wireless connection methods. For example, in the business scenario of insurance operations in the financial insurance industry, the aforementioned operational process may include auto insurance operations, life insurance operations, accident insurance operations, etc. The aforementioned information query request is a request triggered by the target agent during the operational process to obtain customer information of the target customer needed for the corresponding current customer contact stage.

[0087] Step S202: If yes, obtain the job feature information corresponding to the job process based on the preset information type.

[0088] In this embodiment, the aforementioned information types may specifically include agent operation mode, agent operation, agent experience, agent profile, customer profile, contact stage information, etc. Correspondingly, the aforementioned operation feature information corresponding to the operation process includes at least the agent's current operation mode, agent's current operation information, agent's experience, current operation steps, agent profile, customer profile, contact stage information, etc.

[0089] Step S203: Input the job feature information into a pre-trained information prediction model, and use the information prediction model to predict the job feature information to generate prediction probability values ​​for the job feature information corresponding to various classification label information.

[0090] In this embodiment, the aforementioned information prediction model is generated by training and filtering an initial model based on the XGBoost algorithm using pre-collected historical job feature information. The specific generation process of the information prediction model will be described in further detail in subsequent embodiments of this application, and will not be elaborated upon here.

[0091] Step S204: Based on each of the predicted probability values, determine the target classification label information that meets the preset importance conditions from all the classification label information.

[0092] In this embodiment, the specific implementation process of determining the target classification label information that meets the preset importance conditions from all the classification label information based on each of the predicted probability values ​​will be further described in detail in subsequent specific embodiments of this application, and will not be elaborated on here.

[0093] Step S205: Obtain the target query information of the target customer corresponding to the target category tag information.

[0094] In this embodiment, the agent operation includes multiple business steps, and the customer information required to support the operation differs in each step. Customer information for all customers is pre-stored. Specifically, customer information may include basic customer information, customer profile information, vehicle information, historical contact information, historical quote information, historical insurance information, etc. By querying the aforementioned customer information, the specific customer information of the target customer corresponding to the target category tag information can be retrieved, thus obtaining the aforementioned target query information.

[0095] Step S206: Push the target query information to the target agent's work interface.

[0096] In this embodiment, the specific implementation process of pushing the target query information to the target agent's work interface will be described in more detail in subsequent specific embodiments of this application, and will not be elaborated on here.

[0097] This application first determines whether an information query request corresponding to a target customer, triggered by a target agent during the operation, has been received. If so, based on a preset information type, operation feature information corresponding to the operation is obtained. Then, the operation feature information is input into a pre-trained information prediction model, and the operation feature information is predicted by the information prediction model to generate predicted probability values ​​for the operation feature information corresponding to various classification tag information. Subsequently, based on each predicted probability value, target classification tag information that meets preset importance conditions is determined from all the classification tag information. The target query information of the target customer corresponding to the target classification tag information is further obtained. Finally, the target query information is pushed to the target agent's work interface. This application, upon receiving an information query request corresponding to a target customer triggered by a target agent during a task, automatically acquires task feature information corresponding to the task process. Then, based on a pre-trained information prediction model, it performs predictive processing on the task feature information, generating predicted probability values ​​for various category tag information corresponding to the task feature information. Based on the obtained predicted probability values, it determines the target category tag information that meets the importance criteria, thereby acquiring the target customer's target query information corresponding to the target category tag information. Finally, it pushes the target query information to the target agent's work interface to satisfy the target agent's information access needs. The use of the information prediction model accurately predicts the target query information corresponding to the target customer, ensuring the accuracy of the generated target query information. It also effectively reduces the time spent by the target agent searching for query information, improving the processing efficiency of the target agent in obtaining effective customer information, thus enhancing the target agent's work efficiency and user experience.

[0098] In some alternative implementations, step S204 includes the following steps:

[0099] Obtain the preset probability threshold.

[0100] In this embodiment, the value of the above probability threshold is not specifically limited and can be set according to actual usage requirements.

[0101] Each of the predicted probability values ​​is compared with the probability threshold, and a specified probability threshold greater than the probability threshold is selected from all the probability thresholds.

[0102] In this embodiment, the specified probability threshold can be obtained by comparing each of the predicted probability values ​​with the probability threshold to filter out the probability thresholds that are greater than the probability threshold from all the probability thresholds.

[0103] Select the specified category label information that corresponds to the specified probability threshold from all the category label information.

[0104] In this embodiment, the specified classification label information is obtained by filtering out the classification label information corresponding to the specified probability threshold from all the classification label information.

[0105] The specified category label information is used as the target category label information.

[0106] This application obtains a preset probability threshold; then compares each predicted probability value with the probability threshold, and filters out a specified probability threshold greater than the predicted probability threshold from all the probability thresholds; then filters out specified category label information corresponding to the specified probability threshold from all the category label information; subsequently, the specified category label information is used as the target category label information. This application ensures the accuracy of the generated target category label information by filtering out the specified category label information corresponding to the specified probability threshold with a probability threshold greater than the predicted probability threshold from all the category label information.

[0107] In some optional implementations of this embodiment, step S206 includes the following steps:

[0108] Obtain the target probability value corresponding to each of the target classification label information.

[0109] In this embodiment, the target probability value corresponding to each target classification label information can be extracted from the predicted probability values ​​of the job feature information output by the information prediction model corresponding to various classification label information.

[0110] Information priority is generated for each target query information based on the target probability value.

[0111] In this embodiment, the higher the probability value of the category tag information, the higher the importance of the query information corresponding to the category tag information, and thus the higher the priority of the information corresponding to the category tag information.

[0112] Get the preset display page.

[0113] In this embodiment, the aforementioned display page is a pre-built page used to push information to agents.

[0114] Based on the information priority of each target query information, the display position of each target query information is determined on the display page.

[0115] In this embodiment, query information with different information priorities can be pre-set on the display page according to actual information display needs. The higher the information priority of the query information, the higher its display position on the display page.

[0116] Based on the display position of each of the target query information, the target query information is filled into the display page accordingly to obtain the target display page.

[0117] In this embodiment, the target display page can be obtained by filling each of the target query information into the corresponding display positions on the display page.

[0118] The target display page is pushed to the target agent's work interface.

[0119] In this embodiment, the target display page can be pushed to the target agent's work interface by obtaining the communication information of the target agent's work terminal.

[0120] This application obtains target probability values ​​corresponding to each of the target category label information; then generates information priorities for each of the target query information based on the target probability values; subsequently obtains a preset display page; then determines the display position of each of the target query information on the display page based on the information priorities of each of the target query information; further, based on the display position of each of the target query information, fills each of the target query information into the display page accordingly to obtain the target display page; finally, pushes the target display page to the working interface of the target agent. This application generates information priorities for each target query based on the target probability value corresponding to each target category label information. Based on these priorities, it determines the display position of each target query on a preset display page, then fills the corresponding target query information into the display page and pushes the resulting target display page to the target agent's work interface. This achieves the orderly display of obtained target query information on the target agent's work interface according to information priority, satisfying the target agent's information retrieval needs. It effectively reduces the time spent by target agents searching for information, improves the efficiency of target agents in obtaining effective customer information, and enhances the work efficiency and user experience of target agents.

[0121] In some alternative implementations, prior to step S203, the electronic device may also perform the following steps:

[0122] Obtain pre-collected historical operation feature information.

[0123] In this embodiment, the above refers to historical operational characteristic information generated or used by agents and customers during business communications within a pre-collected time period. This historical operational characteristic information includes at least the agent's work mode, agent's operation information, agent's experience, work process steps, agent profile, customer profile, contact stage information, and customer information. The value of the pre-collected time period is not specifically limited and can be set according to actual usage needs; for example, it can be set to the past six months. Furthermore, the agent information such as the agent's work mode, experience, and contact stage information in the historical operational characteristic information can be generated by pre-built tracking systems that track all agent operations (e.g., functions viewed by the agent, data copied) and analyze the collected tracking data.

[0124] The historical operation feature information is preprocessed to obtain corresponding sample data.

[0125] In this embodiment, the specific implementation process of preprocessing the historical operation feature information to obtain the corresponding sample data will be further described in detail in subsequent specific embodiments of this application, and will not be elaborated on here.

[0126] The sample data is divided into a training set and a test set.

[0127] In this embodiment, the sample data can be randomly divided according to a preset division ratio to obtain corresponding training and test sets. The value of the division ratio is not specifically limited and can be set according to actual usage requirements; for example, it can be set to 7:3.

[0128] Multiple initial models were constructed based on the XGBoost algorithm.

[0129] In this embodiment, the first initial model is built based on the XGBoost algorithm. XGBoost (Extreme Gradient Boosting) is an efficient gradient boosting decision tree algorithm. It improves upon the original GBDT, significantly enhancing model performance. As a forward additive model, its core principle is ensemble thinking—Boosting—which integrates multiple weak learners into a single strong learner. This involves multiple trees making decisions together, with each tree's result being the difference between the target value and the predictions of all previous trees. The final result is obtained by summing all the results, thus improving the overall model performance. XGBoost consists of multiple CART (Classification and Regression Trees), enabling it to handle classification and regression problems. The number of models in the first initial model is not specifically limited and can be set according to actual usage requirements.

[0130] The training set is used to train each of the first initial models to obtain multiple trained second initial models.

[0131] In this embodiment, a first initial model is trained using a training set to learn the association between the feature information contained in the training set and the preset classification label information. This leads to the training of a second initial model, which is used to predict the importance of various customer information needed by the agent at the current stage of their work. The specific training process of the first initial model can refer to existing XGBoost training methods and will not be elaborated upon here.

[0132] Each of the second initial models is used to perform prediction processing on the test set to obtain the prediction results output by each of the second initial models.

[0133] In this embodiment, the prediction results output by each of the second initial models can be obtained by inputting the test set into each of the second initial models respectively.

[0134] Based on the prediction results output by each of the second initial models, model screening is performed on each of the second initial models to determine the information prediction model from all the second initial models.

[0135] In this embodiment, the specific implementation process of selecting the information prediction model from all the second initial models based on the prediction results output by each second initial model will be described in more detail in subsequent specific embodiments of this application, and will not be elaborated on here.

[0136] This application obtains pre-collected historical task feature information; then preprocesses the historical task feature information to obtain corresponding sample data; divides the sample data into training and testing sets; then constructs multiple first initial models based on the XGBoost algorithm; subsequently, the training set is used to train each of the first initial models to obtain multiple trained second initial models; further, each of the second initial models is used to predict the test set to obtain the prediction results output by each of the second initial models; finally, the prediction results output by each of the second initial models are used to filter the second initial models and determine the information prediction model from all the second initial models. This application effectively ensures the accuracy of the generated information prediction model by using pre-collected historical task feature information as sample data, then using the training set in the sample data to train the first initial models constructed based on the XGBoost algorithm to obtain second initial models, and using the testing set in the sample data to filter the second initial models to determine the final information prediction model. This facilitates the subsequent accurate prediction processing of the task feature information using the information prediction model to generate prediction probability values ​​for task feature information corresponding to various classification label information, thus ensuring the accuracy of the generated prediction information.

[0137] In some optional implementations, the preprocessing of the historical job feature information to obtain corresponding sample data includes the following steps:

[0138] The historical job feature information is cleaned to obtain the corresponding first job feature information.

[0139] In this embodiment, the above-mentioned data cleaning process refers to cleaning the data in the historical operation feature information that has lost information or abnormal data, so as to obtain the above-mentioned first operation feature information.

[0140] The first job feature information is converted into a format to obtain the corresponding second job feature information.

[0141] In this embodiment, the first text can be formatted using a data conversion tool to obtain data that conforms to the data format that the model can process, namely the above-mentioned second job feature information.

[0142] The second job feature information is used as the sample data.

[0143] This application cleanses the historical job feature information to obtain corresponding first job feature information; then, it performs format conversion on the first job feature information to obtain corresponding second job feature information; subsequently, the second job feature information is used as the sample data. By performing data cleaning and format conversion on the obtained historical job feature information, this application can quickly obtain sample data with a data format that conforms to the model's readable and processing requirements, effectively ensuring the smooth progress of the subsequent model training process for obtaining the information prediction model using the sample data.

[0144] In some optional implementations of this embodiment, the step of filtering the second initial models based on the prediction results output by each second initial model to determine the information prediction model from all the second initial models includes the following steps:

[0145] Based on the prediction results output by each of the second initial models, the second initial models are evaluated to obtain the processing effect value of each of the second initial models.

[0146] In this embodiment, the above-mentioned processing effect value may include any one of the following evaluation values: recall rate, precision rate, and F1 score.

[0147] Numerical analysis was performed on all the processing effect values, and the third initial model with the highest processing effect value was selected from all the second initial models.

[0148] In this embodiment, the third initial model with the highest processing effect value refers to the model with the best prediction effect among all the second initial models.

[0149] The third initial model is used as the information prediction model.

[0150] This application evaluates each of the second initial models based on the prediction results output by each second initial model to obtain the processing effect value of each second initial model; then, it performs numerical analysis on all the processing effect values ​​and selects the third initial model with the highest processing effect value from all the second initial models; subsequently, the third initial model is used as the information prediction model. This application evaluates each of the second initial models based on the prediction results output by each second initial model to obtain the processing effect value of each second initial model, and then selects the third initial model with the highest processing effect value from all the second initial models as the final information prediction model. Since this third initial model has the best prediction effect among all the second initial models, it effectively ensures the accuracy of the generated information prediction model, which is beneficial for subsequent use of the information prediction model to accurately predict and process the job feature information, thereby generating prediction probability values ​​for the job feature information corresponding to various classification label information, and ensuring the accuracy of the generated prediction information.

[0151] In some optional implementations of this embodiment, after the step of performing model filtering on each of the second initial models based on the prediction results output by each of the second initial models to determine the information prediction model from all the second initial models, the electronic device may further perform the following steps:

[0152] The information prediction model is then published.

[0153] In this embodiment, after determining the information prediction model from all the second initial models, the information prediction model will be further published and an API interface will be provided to the outside world so that the API interface can be called in the future.

[0154] After the information prediction model is published, it is stored.

[0155] In this embodiment, the storage method for the information prediction model is not specifically limited and can be set according to actual usage needs. For example, storage methods such as blockchain storage, database storage, and cloud storage can be used.

[0156] Obtain a fourth initial model from all the second initial models except for the information prediction model.

[0157] In this embodiment, the four initial models mentioned above refer to all the other models in all the second initial models except for the information prediction model.

[0158] The fourth initial model is then deleted.

[0159] In this embodiment, by deleting the fourth initial model, the storage space of the electronic device can be effectively reduced.

[0160] This application involves publishing the information prediction model; storing the information prediction model after publishing it; then obtaining a fourth initial model from all the second initial models other than the information prediction model; and subsequently deleting the fourth initial model. After determining the information prediction model from all the second initial models, this application intelligently publishes the information prediction model to ensure smooth subsequent calls to it. The information prediction model is also stored to ensure its security. Furthermore, the fourth initial model from all the second initial models other than the information prediction model is deleted to remove invalid models, effectively improving the intelligence of model storage, reducing the storage space of electronic devices, and avoiding impact on the normal operation of electronic devices.

[0161] It should be understood that the sequence number of each step in the above embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0162] It should be emphasized that, to further ensure the privacy and security of the aforementioned target query information, the target query information can also be stored in a blockchain node.

[0163] The blockchain referred to in this application is a novel application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanisms, and encryption algorithms. Essentially, a blockchain is a decentralized database, a chain of data blocks linked together using cryptographic methods. Each data block contains information about a batch of network transactions, used to verify the validity of the information (anti-counterfeiting) and generate the next block. A blockchain can include an underlying blockchain platform, a platform product service layer, and an application service layer.

[0164] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0165] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0166] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by instructing related hardware with computer-readable instructions. These computer-readable instructions can be stored in a computer-readable storage medium. When executed, the program can include the processes of the embodiments of the above methods. The aforementioned storage medium can be a non-volatile storage medium such as a magnetic disk, optical disk, or read-only memory (ROM), or random access memory (RAM).

[0167] It should be understood that although the steps in the flowcharts of the accompanying figures are shown sequentially as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the accompanying figures may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the sub-steps or stages of other steps.

[0168] Further reference Figure 3 As a response to the above Figure 2 To implement the method shown, this application provides an embodiment of an artificial intelligence-based information push device, which is similar to... Figure 2 Corresponding to the method embodiments shown, this device can be specifically applied to various electronic devices.

[0169] like Figure 3 As shown, the AI-based information push device 300 described in this embodiment includes: a loading module 301, a calling module 302, a collection module 303, a generation module 304, and a processing module 305. Wherein:

[0170] The judgment module 301 is used to determine whether an information query request corresponding to the target customer has been received by the target agent during the operation process;

[0171] The first acquisition module 302 is used to acquire, if so, job feature information corresponding to the job process based on a preset information type;

[0172] Prediction module 303 is used to input the job feature information into a pre-trained information prediction model, and to perform prediction processing on the job feature information through the information prediction model to generate prediction probability values ​​of the job feature information corresponding to various classification label information.

[0173] The first determining module 304 is used to determine, based on each of the predicted probability values, target classification label information that meets the preset importance conditions from all the classification label information;

[0174] The second acquisition module 305 is used to acquire the target query information of the target customer corresponding to the target category label information;

[0175] The push module 306 is used to push the target query information to the target agent's work interface.

[0176] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based information push method in the aforementioned implementation method, and will not be repeated here.

[0177] In some optional implementations of this embodiment, the first determining module 304 includes:

[0178] The first acquisition submodule is used to acquire a preset probability threshold;

[0179] The first filtering submodule is used to compare each of the predicted probability values ​​with the probability threshold respectively, and filter out a specified probability threshold that is greater than the probability threshold from all the probability thresholds;

[0180] The second filtering submodule is used to filter out the specified category label information that corresponds to the specified probability threshold from all the category label information;

[0181] The determination submodule is used to use the specified category label information as the target category label information.

[0182] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based information push method in the aforementioned implementation method, and will not be repeated here.

[0183] In some optional implementations of this embodiment, the push module 306 includes:

[0184] The first acquisition submodule is used to acquire the target probability value corresponding to each of the target classification label information;

[0185] Based on the target probability value, generate the information priority of each target query information;

[0186] Get the preset display page;

[0187] Based on the information priority of each of the target query information, the display position of each of the target query information is determined on the display page;

[0188] Based on the display position of each of the target query information, each of the target query information is filled into the display page accordingly to obtain the target display page;

[0189] The target display page is pushed to the target agent's work interface.

[0190] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based information push method in the aforementioned implementation method, and will not be repeated here.

[0191] In some optional implementations of this embodiment, the AI-based information push device further includes:

[0192] The third acquisition module is used to acquire pre-collected historical operation feature information;

[0193] The preprocessing module is used to preprocess the historical operation feature information to obtain corresponding sample data;

[0194] A partitioning module is used to partition the sample data into a training set and a test set;

[0195] The building module is used to construct multiple corresponding first initial models based on the XGBoost algorithm;

[0196] The training module is used to train each of the first initial models using the training set to obtain multiple trained second initial models.

[0197] The processing module is used to perform prediction processing on the test set using each of the second initial models to obtain the prediction results output by each of the second initial models;

[0198] The second determining module is used to perform model screening on each of the second initial models based on the prediction results output by each of the second initial models, and to determine the information prediction model from all the second initial models.

[0199] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based information push method in the aforementioned implementation method, and will not be repeated here.

[0200] In some optional implementations of this embodiment, the preprocessing module includes:

[0201] The first processing submodule is used to perform data cleaning processing on the historical job feature information to obtain the corresponding first job feature information;

[0202] The second processing submodule is used to perform format conversion processing on the first job feature information to obtain the corresponding second job feature information;

[0203] The first determining submodule is used to use the second job feature information as the sample data.

[0204] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based information push method in the aforementioned implementation method, and will not be repeated here.

[0205] In some optional implementations of this embodiment, the second determining module includes:

[0206] The evaluation submodule is used to evaluate each of the second initial models based on the prediction results output by each of the second initial models, and to obtain the processing effect value of each of the second initial models.

[0207] The analysis submodule is used to perform numerical analysis on all the processing effect values ​​and to select the third initial model with the highest processing effect value from all the second initial models.

[0208] The second determining submodule is used to use the third initial model as the information prediction model.

[0209] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based information push method in the aforementioned implementation method, and will not be repeated here.

[0210] In some optional implementations of this embodiment, the AI-based information push device further includes:

[0211] The publishing module is used to process the publishing of the information prediction model;

[0212] The storage module is used to store the information prediction model after the publishing process of the information prediction model is completed.

[0213] The fourth acquisition module is used to acquire all the fourth initial models in the second initial models other than the information prediction model;

[0214] The deletion module is used to delete the fourth initial model.

[0215] In this embodiment, the operations performed by the above modules or units correspond one-to-one with the steps of the artificial intelligence-based information push method in the aforementioned implementation method, and will not be repeated here.

[0216] To address the aforementioned technical problems, embodiments of this application also provide a computer device. Please refer to [link / reference needed] for details. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0217] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected via a system bus. It should be noted that only the computer device 4 with components 41-43 is shown in the figure; however, it should be understood that it is not required to implement all the shown components, and more or fewer components can be implemented alternatively. Those skilled in the art will understand that the computer device described here is a device capable of automatically performing numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes, but is not limited to, microprocessors, application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), digital signal processors (DSPs), embedded devices, etc.

[0218] The computer device can be a desktop computer, laptop, handheld computer, or cloud server, etc. The computer device can interact with the user via a keyboard, mouse, remote control, touchpad, or voice control.

[0219] The memory 41 includes at least one type of readable storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 41 may be an internal storage unit of the computer device 4, such as the hard disk or memory of the computer device 4. In other embodiments, the memory 41 may also be an external storage device of the computer device 4, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the computer device 4. Of course, the memory 41 may also include both the internal storage unit and its external storage device of the computer device 4. In this embodiment, the memory 41 is typically used to store the operating system and various application software installed on the computer device 4, such as computer-readable instructions for information push methods based on artificial intelligence. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or will be output.

[0220] In some embodiments, the processor 42 may be a central processing unit (CPU), controller, microcontroller, microprocessor, or other data processing chip. The processor 42 is typically used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to execute computer-readable instructions stored in the memory 41 or to process data, for example, to execute computer-readable instructions for the AI-based information push method.

[0221] The network interface 43 may include a wireless network interface or a wired network interface, which is typically used to establish communication connections between the computer device 4 and other electronic devices.

[0222] Compared with the prior art, the embodiments of this application have the following main advantages:

[0223] In this embodiment, when a target agent receives an information query request corresponding to a target customer triggered during the operation, the system automatically acquires the operation feature information corresponding to the operation process. Then, based on a pre-trained information prediction model, the system predicts the operation feature information to generate predicted probability values ​​for various classification tag information. Based on these predicted probability values, the system determines the target classification tag information that meets the importance criteria. This allows the system to obtain the target customer's query information corresponding to the target classification tag information. Finally, the target query information is pushed to the target agent's work interface to meet the agent's information access needs. The use of the information prediction model accurately predicts the target query information corresponding to the target customer, ensuring the accuracy of the generated query information. It also effectively reduces the time spent by the target agent searching for query information, improving the agent's efficiency in obtaining effective customer information, thus enhancing the agent's work efficiency and user experience.

[0224] This application also provides another embodiment, namely, providing a computer-readable storage medium storing computer-readable instructions that can be executed by at least one processor to cause the at least one processor to perform the steps of the artificial intelligence-based information push method described above.

[0225] Compared with the prior art, the embodiments of this application have the following main advantages:

[0226] In this embodiment, when a target agent receives an information query request corresponding to a target customer triggered during the operation, the system automatically acquires the operation feature information corresponding to the operation process. Then, based on a pre-trained information prediction model, the system predicts the operation feature information to generate predicted probability values ​​for various classification tag information. Based on these predicted probability values, the system determines the target classification tag information that meets the importance criteria. This allows the system to obtain the target customer's query information corresponding to the target classification tag information. Finally, the target query information is pushed to the target agent's work interface to meet the agent's information access needs. The use of the information prediction model accurately predicts the target query information corresponding to the target customer, ensuring the accuracy of the generated query information. It also effectively reduces the time spent by the target agent searching for query information, improving the agent's efficiency in obtaining effective customer information, thus enhancing the agent's work efficiency and user experience.

[0227] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes several instructions to cause a terminal device (which may be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0228] Obviously, the embodiments described above are only some embodiments of this application, not all embodiments. The accompanying drawings show preferred embodiments of this application, but do not limit the patent scope of this application. This application can be implemented in many different forms; rather, the purpose of providing these embodiments is to provide a more thorough and comprehensive understanding of the disclosure of this application. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing specific embodiments, or make equivalent substitutions for some of the technical features. Any equivalent structures made using the content of this application's specification and drawings, directly or indirectly applied to other related technical fields, are similarly within the scope of patent protection of this application.

Claims

1. An information push method based on artificial intelligence, characterized in that, Includes the following steps: Determine whether an information query request corresponding to the target customer has been received by the target agent during the operation process; If so, based on the preset information type, obtain the operation feature information corresponding to the operation process; wherein, the operation feature information corresponding to the operation process includes the current operation mode of the agent, the current operation information of the agent, the agent's experience, the current process steps of the operation, agent profile, customer profile, and contact stage information. The task feature information is input into a pre-trained information prediction model, and the information prediction model is used to predict the task feature information to generate prediction probability values ​​of the task feature information corresponding to various classification label information. Based on the predicted probability values, target classification label information that meets the preset importance conditions is determined from all the classification label information; Obtain the target query information of the target customer corresponding to the target category tag information; The target query information is pushed to the target agent's work interface; The step of inputting the job feature information into a pre-trained information prediction model, performing prediction processing on the job feature information through the information prediction model, and generating predicted probability values ​​for the job feature information corresponding to various classification label information further includes: Obtain pre-collected historical operation feature information; The historical operation feature information is preprocessed to obtain corresponding sample data; The sample data is divided into a training set and a test set; Multiple initial models were constructed based on the XGBoost algorithm. The training set is used to train each of the first initial models to obtain multiple trained second initial models. Each of the second initial models is used to perform prediction processing on the test set to obtain the prediction results output by each of the second initial models; Based on the prediction results output by each of the second initial models, model screening is performed on each of the second initial models to determine the information prediction model from all the second initial models.

2. The information push method based on artificial intelligence according to claim 1, characterized in that, The step of determining the target classification label information that meets the preset importance condition from all the classification label information based on each of the predicted probability values ​​specifically includes: Obtain the preset probability threshold; Each of the predicted probability values ​​is compared with the probability threshold, and a specified probability threshold greater than the probability threshold is selected from all the probability thresholds. Filter out the specified category label information that corresponds to the specified probability threshold from all the category label information; The specified category label information is used as the target category label information.

3. The information push method based on artificial intelligence according to claim 1, characterized in that, The step of pushing the target query information to the target agent's work interface specifically includes: Obtain the target probability value corresponding to each of the target classification label information; Based on the target probability value, generate the information priority of each target query information; Get the preset display page; Based on the information priority of each of the target query information, the display position of each of the target query information is determined on the display page; Based on the display position of each of the target query information, each of the target query information is filled into the display page accordingly to obtain the target display page; The target display page is pushed to the target agent's work interface.

4. The information push method based on artificial intelligence according to claim 1, characterized in that, The step of preprocessing the historical operation feature information to obtain corresponding sample data specifically includes: The historical job feature information is cleaned to obtain the corresponding first job feature information; The first job feature information is converted to obtain the corresponding second job feature information; The second job feature information is used as the sample data.

5. The information push method based on artificial intelligence according to claim 1, characterized in that, The step of filtering the second initial models based on the prediction results output by each second initial model, and determining the information prediction model from all the second initial models, specifically includes: Based on the prediction results output by each of the second initial models, the second initial models are evaluated to obtain the processing effect value of each of the second initial models; Numerical analysis was performed on all the processing effect values, and the third initial model with the highest processing effect value was selected from all the second initial models. The third initial model is used as the information prediction model.

6. The information push method based on artificial intelligence according to claim 1, characterized in that, After the step of filtering each of the second initial models based on the prediction results output by each of the second initial models to determine the information prediction model from all the second initial models, the method further includes: The information prediction model is then processed for publication. After completing the publishing process of the information prediction model, the information prediction model is stored. Obtain a fourth initial model from all the second initial models except for the information prediction model; The fourth initial model is then deleted.

7. An information push device based on artificial intelligence, characterized in that, include: The judgment module is used to determine whether an information query request corresponding to the target customer has been received by the target agent during the operation. The first acquisition module is used to acquire, if so, job feature information corresponding to the job process based on a preset information type; wherein, the job feature information corresponding to the job process includes the agent's current job mode, agent's current operation information, agent's experience, current job process steps, agent profile, customer profile, and contact stage information. The prediction module is used to input the job feature information into a pre-trained information prediction model, and to perform prediction processing on the job feature information through the information prediction model to generate prediction probability values ​​of the job feature information corresponding to various classification label information. The first determining module is used to determine, based on each of the predicted probability values, target classification label information that meets the preset importance conditions from all the classification label information; The second acquisition module is used to acquire the target query information of the target customer corresponding to the target category label information; The push module is used to push the target query information to the target agent's work interface; Among them, the information push device based on artificial intelligence also includes: The third acquisition module is used to acquire pre-collected historical operation feature information; The preprocessing module is used to preprocess the historical operation feature information to obtain corresponding sample data; A partitioning module is used to partition the sample data into a training set and a test set; The building module is used to construct multiple corresponding first initial models based on the XGBoost algorithm; The training module is used to train each of the first initial models using the training set to obtain multiple trained second initial models. The processing module is used to perform prediction processing on the test set using each of the second initial models to obtain the prediction results output by each of the second initial models; The second determining module is used to perform model screening on each of the second initial models based on the prediction results output by each of the second initial models, and to determine the information prediction model from all the second initial models.

8. A computer device, comprising a memory and a processor, wherein the memory stores computer-readable instructions, and the processor, when executing the computer-readable instructions, implements the steps of the information push method based on artificial intelligence as described in any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the information push method based on artificial intelligence as described in any one of claims 1 to 6.

Images