Training method and device, electronic equipment and storage medium

Abstract

The application discloses a training method and device, electronic equipment and storage medium, belongs to the technical field of artificial intelligence, and can be applied to a financial science and technology scene. The method comprises the following steps: acquiring learning data of a training object; determining the similarity between the learning data of the training object and the learning data of each candidate object in a plurality of candidate objects, and determining a candidate object satisfying a preset similarity condition as a similar object of the training object; determining a classification label of the training object according to attribute data of the training object; generating a training path corresponding to the training object according to course information learned by the similar object and the classification label, wherein the training path comprises at least one training course; and pushing the training course to the training object in the order of the training courses in the training path. The embodiment of the application can improve the training efficiency of the training object.

Description

Technical Field

[0001] This application belongs to the field of artificial intelligence technology, and in particular relates to a training method, device, electronic device and storage medium. Background Technology

[0002] With the rapid development of property insurance, the professional competence of insurance agents directly impacts the service quality and business efficiency of insurance institutions. Therefore, training insurance agents has become a crucial aspect of insurance institutions' business operations. Currently, training content is typically provided to different insurance agents through either in-person lectures or online pre-recorded courses.

[0003] However, the training content provided by the above-mentioned training methods is highly homogeneous and fails to take into account the business characteristics and individual knowledge gaps of insurance agents. This results in a mismatch between the training content and the actual training needs of insurance agents, poor training relevance, and low training efficiency. Summary of the Invention

[0004] This application provides a training method, apparatus, electronic device, and storage medium that can improve the training efficiency for trainees.

[0005] To achieve the above objectives, a first aspect of this application provides a training method, the method comprising: Obtain learning data from trainees; Determine the similarity between the learning data of the training object and the learning data of each candidate object among multiple candidate objects, and determine the candidate objects whose similarity satisfies the preset similarity conditions as similar objects of the training object; Based on the attribute data of the training subjects, determine the classification labels of the training subjects; Based on the course information learned by the similar objects and the classification tags, a training path corresponding to the training object is generated, wherein the training path includes at least one training course. The training courses are pushed to the trainees in the order of the training courses in the training path.

[0006] In some embodiments, the method further includes: The target data is obtained by web crawling, including at least one of the following: industry policy data, insurance premium adjustment data, and claims public opinion data. Extract key information from the target data to obtain industry dynamic data; The course content of each training course is updated based on the aforementioned industry dynamic data.

[0007] In some embodiments, before determining the classification label of the training object based on the attribute data of the training object, the method further includes: Obtain the business data and customer profile data of the training subjects; The attribute data of the training subject is determined based on at least one of the business data and the customer profile data.

[0008] In some embodiments, the similarity between the learning data of the training subjects and the learning data of the candidate subjects is determined in the following manner: For each candidate object, a collaborative filtering algorithm is used to calculate the similarity between the learning data of the training object and the learning data of the candidate object.

[0009] In some embodiments, after generating the training path corresponding to the training object, the method further includes at least one of the following: Adjust the order of each training course in the training path according to the learning progress of the trainees; In response to the training subject's test accuracy rate for the first course in the training path being lower than a preset threshold, a supplementary course associated with the first course is added to the training path.

[0010] In some embodiments, after generating the training path corresponding to the training object based on the course information learned by the similar objects and the classification tags, the method further includes: Displaying a three-dimensional virtual business scene, wherein the three-dimensional virtual business scene includes dynamic interactive elements, and the dynamic interactive elements include at least one of virtual customer status change elements, emergency situation elements, and error reminder elements; In response to the training operation performed by the training subject in the three-dimensional virtual business scenario, the dynamic interactive elements in the three-dimensional virtual business scenario are updated according to the training operation; The communication voice between the training subject and the virtual customer in the three-dimensional virtual business scenario is identified by speech recognition to obtain the communication text. Semantic analysis and sentiment analysis are performed on the communication text to generate speech optimization suggestions corresponding to the training operation. Obtain training data of the training subject performing the training operation, wherein the training data includes at least one of the communication text, the completion time of the training operation, and the document filling result corresponding to the training operation; Based on the training data, evaluate the training effectiveness of the trainees.

[0011] In some embodiments, evaluating the training effectiveness of the trainees based on the training data includes: Construct an evaluation hierarchy, wherein the evaluation hierarchy includes an objective layer, a criterion layer, and an indicator layer under the criterion layer; The evaluation indicators at the same level in the evaluation hierarchy are compared pairwise to obtain a judgment matrix; The weights of each evaluation index are determined based on the judgment matrix, and the consistency of the judgment matrix is ​​checked. Based on the training data, a fuzzy relation matrix is ​​determined, wherein the elements in the fuzzy relation matrix are used to characterize the evaluation level of the training object corresponding to the evaluation index; If the judgment matrix passes the consistency test, then the weight vector composed of the weights of each evaluation index is combined with the fuzzy relation matrix to perform a fuzzy synthesis operation to obtain the ability evaluation value of the training object. Based on the capability assessment values, a capability radar chart of the trainee is generated, which is used to characterize the training effect of the trainee.

[0012] To achieve the above objectives, a second aspect of this application provides a training apparatus, the apparatus comprising: The acquisition module is used to acquire the learning data of the trainees; The first determining module is used to determine the similarity between the learning data of the training object and the learning data of each candidate object among multiple candidate objects, and to determine the candidate objects whose similarity satisfies the preset similarity conditions as similar objects of the training object. The second determining module is used to determine the classification label of the training object based on the attribute data of the training object; The generation module is used to generate a training path corresponding to the training object based on the course information learned by the similar objects and the classification tags, wherein the training path includes at least one training course. The push module is used to push the training courses to the trainees in the order of the training courses in the training path.

[0013] To achieve the above objectives, a third aspect of this application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the method described in the first aspect.

[0014] To achieve the above objectives, a fourth aspect of the present application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method described in the first aspect.

[0015] The training method, apparatus, electronic device, and storage medium proposed in this application acquire the learning data of the trainees; based on the learning data, identify similar objects with high similarity to the trainees from multiple candidate objects; simultaneously, determine the classification tags of the trainees based on their attribute data; then, combining the course information learned by similar objects and the classification tags, generate a personalized training path for the trainees containing at least one training course, and push the training courses to the trainees according to this training path. In this way, by generating training paths for trainees based on the behavior of similar objects and the trainees' own attribute tags, trainees do not need to manually select from a general course library or passively receive all courses, but directly obtain a combination of courses highly matched to their business scenarios and learning needs, reducing ineffective learning content and repetitive learning time, thereby improving training efficiency. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the application environment for the training method provided in the embodiments of this application; Figure 2 This is an optional flowchart of the training method provided in the embodiments of this application; Figure 3 This is another optional flowchart of the training method provided in the embodiments of this application; Figure 4 This is an optional flowchart of the training method provided in the embodiments of this application; Figure 5 yes Figure 3 Flowchart of step S305; Figure 6 This is a schematic diagram of the structure of the training device provided in the embodiments of this application; Figure 7 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0019] It should be noted that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, and the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0020] 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 belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0021] First, let's analyze some of the terms used in this application: Artificial intelligence (AI) is a new branch of computer science that studies, develops, and applies theories, methods, technologies, and systems to simulate, extend, and expand human intelligence. It aims to understand the essence of intelligence and produce intelligent machines that can react in a way similar to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing, and expert systems. AI can simulate the information processes of human consciousness and thought. Furthermore, AI utilizes digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceiving the environment, acquiring knowledge, and using that knowledge to achieve optimal results.

[0022] For ease of understanding, the following describes some aspects of the embodiments of this application: Currently, the training model for insurance agents in the property insurance industry mainly consists of offline centralized lectures and online recorded courses. Some institutions have introduced artificial intelligence (AI) teaching assistants to help answer questions, but their capabilities are limited, and the existing solutions have the following shortcomings: Firstly, the turnover rate of agents is high. New agents need to undergo a full set of basic training again after joining the company. The organization needs to repeatedly invest human and material resources such as lecturers, venues, and teaching materials. At the same time, the repetitive training process also prolongs the onboarding period for new agents, affecting the efficiency of the organization's business expansion. Secondly, the training content is highly homogenized and fails to take into account the insurance needs of the agent's region, the characteristics of the customer group, and the agent's personal knowledge gaps, resulting in weak training relevance and requiring agents to spend a lot of time learning irrelevant content. Third, the efficiency of practical application is low. Existing training emphasizes theory over practice, lacks real-world scenario simulations, claims case analysis, and customer communication skills training, making it difficult for agents to quickly transform knowledge into business development capabilities. Fourth, the training effect is difficult to quantify and evaluate. There is a lack of real-time tracking of the agent's learning progress, mastery of knowledge points, and practical performance. The organization cannot accurately judge the training effect and it is difficult to develop personalized reinforcement plans. Fifth, the updates and iterations are lagging behind. Property insurance policies, regulatory policies, and market dynamics change frequently, and the existing training content has a long update cycle, resulting in a disconnect between the knowledge learned by agents and actual business, which affects the quality of business development.

[0023] Based on this, embodiments of this application provide a training method, apparatus, equipment, and medium, aiming to improve the training efficiency for trainees.

[0024] The recommended methods, apparatuses, electronic devices, and storage media provided in this application are specifically described through the following embodiments. First, the recommended methods in the embodiments of this application are described.

[0025] 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.

[0026] 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.

[0027] The training method provided in this application relates to the field of artificial intelligence technology. The training method provided in this application can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, etc.; the server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application implementing the training method, but is not limited to the above forms.

[0028] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0029] It should be noted that in all specific embodiments of this application, when processing data related to user identity or characteristics, such as user information, user behavior data, user historical data, and user location information, user permission or consent is obtained first. Furthermore, the collection, use, and processing of this data comply with relevant laws, regulations, and standards. In addition, when embodiments of this application require access to sensitive personal information of users, separate permission or consent from the user is obtained through pop-ups or redirection to confirmation pages. Only after obtaining the user's separate permission or consent is the necessary user-related data required for the proper functioning of these embodiments acquired.

[0030] Figure 1 This is a schematic diagram illustrating the application environment of the training method provided in the embodiments of this application. For example... Figure 1 As shown, the application environment may include terminal device 110, server 120, data source device 130 and network 140, and terminal device 110, server 120 and data source device 130 are connected to each other through network 140.

[0031] The terminal device 110 refers to the device used by the trainees. The terminal device is used to display training content and collect data on the trainees' actions. Specifically, the terminal device 110 can be a VR device, a mobile terminal, a personal computer, a tablet computer, etc. The terminal device 110 is used to display training courses and 3D virtual business scenarios within the training path to the trainees, and to collect data such as the trainees' practical training operations, test answer records, and course viewing behavior.

[0032] Server 120 refers to the device used to generate training paths and push training courses. Server 120 can be configured as a standalone physical server, a server cluster or distributed system consisting of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, and cloud computing. Server 120 can deploy a course library and training models. The course library stores various training courses, and the training models can include collaborative filtering algorithm models and decision tree models. Server 120 generates training paths corresponding to the training subjects based on the learning data reported by terminal device 110 and the data provided by data source device 130, and pushes training courses to the training subjects through terminal device 110.

[0033] Data source device 130 refers to the device that provides data to server 120. Specifically, data source device 130 can be the server where the insurance institution's business system is located, used to provide business data and customer profile data of the training subjects; it can also be the server where the industry data platform is located, used to provide industry policy data, insurance rate adjustment data, and claims public opinion data, etc. Network 140 can be a wired network or a wireless network, such as a mobile communication network or a wireless fidelity network.

[0034] It should be noted that the training method provided in this application embodiment can be executed by server 120, terminal device 110, or jointly executed by terminal device 110 and server 120. This application embodiment does not limit this. For ease of description, the following embodiments use an electronic device as the execution subject. The electronic device can be the aforementioned terminal device 110 or the aforementioned server 120.

[0035] Figure 2 This is an optional flowchart of the training method provided in the embodiments of this application. Figure 2 The method may include, but is not limited to, steps S101 to S105.

[0036] Step S101: Obtain the learning data of the trainees.

[0037] Step S102: Determine the similarity between the learning data of the training object and the learning data of each candidate object among multiple candidate objects, and determine the candidate objects whose similarity meets the preset similarity conditions as similar objects of the training object; Step S103: Determine the classification label of the training object based on the attribute data of the training object; Step S104: Based on the course information learned by the similar objects and the classification tags, generate a training path corresponding to the training object, wherein the training path includes at least one training course. Step S105: Push the training courses to the trainees in the order of the training courses in the training path.

[0038] Steps S101 to S105, as illustrated in this embodiment, involve acquiring the learning data of the training subject; identifying similar objects with high similarity to the training subject from multiple candidate objects based on the learning data; determining the training subject's classification label based on the training subject's attribute data; and then generating a personalized training path for the training subject containing at least one training course by combining the course information learned by the similar objects and the classification label, and pushing the training course to the training subject according to this training path. In this way, generating a training path for the training subject based on the behavior of similar objects and the training subject's own attribute labels allows the training subject to directly obtain a course combination highly matched to their business scenario and learning needs, without having to manually select from a general course library or passively accept all courses. This reduces ineffective learning content and repetitive learning time, thereby improving training efficiency.

[0039] In step S101 of some embodiments, the aforementioned training recipient refers to an individual who needs to receive training. For example, the training recipient can be any individual who needs to improve their business capabilities through training. For instance, in a property insurance agent empowerment training scenario, the aforementioned training recipient could be a newly hired property insurance agent or a senior property insurance agent who needs to strengthen specific business capabilities.

[0040] In some embodiments, the aforementioned learning data refers to data that reflects the learning status of the trainees. Specifically, learning data can be any data that can characterize the courses the trainees have learned and their learning outcomes, including but not limited to at least one of the following: course identifiers, completion rates of each course, test answer records, and course viewing behavior data. For example, the aforementioned learning data could be the trainees' online test answer logs, course viewing behavior, and onboarding assessment scores recorded through their personal learning channels. These data form the trainees' personal knowledge graph, which can then serve as their learning data. The online test answer logs may include the types of incorrect answers and the duration of answering; course viewing behavior may include the number of times key chapters were repeatedly viewed and the marking of skipped content.

[0041] The acquisition of training data can be achieved by electronic devices reading training data from the database corresponding to the personal learning channel, or by electronic devices collecting training data in real time from the learning behavior generated by the training on the front-end interactive page and generating learning data accordingly. This application embodiment is not limited to these methods.

[0042] In step S102 of some embodiments, the candidate object can be another object besides the training object that has accumulated certain learning data. For example, the candidate object can be other property insurance agents within the organization besides the current training object.

[0043] In some embodiments, the aforementioned similar objects refer to candidate objects whose similarity to the learning data of the training object meets a preset similarity condition. The aforementioned similarity is used to characterize the degree of closeness between the training object and the candidate objects in terms of learning preferences. The aforementioned preset similarity condition refers to the conditions used to filter similar objects. Specifically, the preset similarity condition can be a similarity greater than or equal to a preset similarity threshold, or it can be that the similarity is ranked within the top K positions after sorting from high to low, where K can be any set positive integer. For example, the aforementioned preset similarity condition could be a similarity higher than 0.7. In this case, candidate objects whose similarity to the training object's learning data is higher than 0.7 can be identified as similar objects, thereby obtaining a similar agent group corresponding to the training object.

[0044] In some embodiments, the aforementioned determination of the similarity between the training subject's learning data and the learning data of each candidate subject, and the identification of candidate subjects whose similarity meets preset similarity conditions as similar subjects, can be achieved by an electronic device calculating the similarity between the training subject's learning data and the learning data of each candidate subject using a collaborative filtering algorithm, thereby identifying candidate subjects with similar learning preferences to the training subject. For example, a collaborative filtering algorithm can identify agents in the same region and at the same job level who account for more than 70% of car insurance sales, and then identify agents with learning preferences related to "rapid accident scene damage assessment" and whose similarity to the training subject is higher than 0.7 as similar subjects to the training subject.

[0045] In step S103 of some embodiments, the aforementioned attribute data refers to data used to describe the characteristics of the training subject. Specifically, the attribute data can be any data that can characterize the training subject's region, job level, business direction, and skill gaps, including but not limited to at least one of the training subject's business data and customer profile data. For example, the aforementioned attribute data can be sales data from agents and customer profile data. Among them, sales data can include the proportion of policy types, customer surrender rates, and the number of claims assistance, while customer profile data can include customer age stratification, occupational distribution, and insurance preferences.

[0046] In some embodiments, the aforementioned classification tags refer to tags used to classify training subjects in multiple dimensions. Specifically, the classification tags can be a combination of tags that characterize training subjects from at least one of the following dimensions: job level, business, region, and skill gaps. For example, the aforementioned classification tags could be "Newcomer - Auto Insurance - First-tier City - High Customer Complaint Rate" or "Experienced - Corporate Property Insurance - Second-tier City - Insufficient Experience in Serving High-Net-Worth Customers".

[0047] In some embodiments, determining the classification labels based on the attribute data of the trainees can be achieved through a decision tree model. This model uses the trainees' attribute data as input to perform multi-dimensional label classification and outputs the aforementioned classification labels. In this way, the classification labels can accurately characterize the trainees' own attribute features, providing a basis for generating personalized training paths.

[0048] In step S104 of some embodiments, the course information learned by similar objects refers to relevant information about courses that similar objects have already learned or frequently learn. Specifically, the course information can be any course identifier that can characterize the learning preferences of similar objects and its learning frequency information. For example, the course information can be courses frequently learned by similar agent groups, such as "the 'Rapid Damage Assessment at Accident Scene' course, which is generally prioritized by agents in the same region, at the same job level, and accounting for more than 70% of car insurance sales."

[0049] In some embodiments, the training path described above refers to a learning path planned for trainees, including at least one training course. Specifically, the training path can be a set of courses composed of at least one of core required courses, personalized elective courses, and real-time updated micro-courses. For example, the training path described above can be a training path that includes "core required courses, personalized elective courses, and real-time updated micro-courses".

[0050] In some embodiments, the above-mentioned generation of training paths based on course information and classification tags learned by similar objects can be achieved by using frequently learned courses by similar objects as candidate courses, and then filtering and organizing the candidate courses in conjunction with the classification tags of the training objects to generate training paths. For example, for training objects who are new agents in first-tier cities with concentrated auto insurance business and a high number of new energy vehicles, the system will automatically push a course package containing "Interpretation of New Energy Vehicle Insurance Terms (Required)", "Sales Scripts for Charging Pile Liability Insurance (Optional)", and "Common Misconceptions in New Energy Vehicle Damage Assessment (Real-time Micro-course)" as the corresponding training path for the training objects.

[0051] In step S105 of some embodiments, after determining the training path, the electronic device can push each training course to the trainee in the order of the training courses in the training path. For example, the electronic device can display each training course in the order of the training path on the front-end interactive page where the trainee logs in, or it can send the access links of each training course to the trainee's terminal device by message push, so that the trainee can complete the learning according to the training path.

[0052] In this embodiment, the learning data of the trainees is acquired; based on the learning data, similar objects with high similarity to the trainees are identified from multiple candidate objects; simultaneously, classification tags for the trainees are determined based on their attribute data; then, combining the course information learned by similar objects and the classification tags, a personalized training path containing at least one training course is generated for the trainees, and training is conducted according to this path. Thus, by generating training paths for trainees based on the behavior of similar objects and the trainees' own attribute tags, trainees do not need to manually select from a general course library or passively accept all courses. Instead, they directly obtain a course combination highly matched to their business scenarios and learning needs, reducing ineffective learning content and repetitive learning time, thereby improving training efficiency.

[0053] Please see Figure 3 In some embodiments, the method further includes steps S201 to S203: Step S201: Use a web crawler to retrieve target data, which includes at least one of the following: industry policy data, insurance premium rate adjustment data, and claims public opinion data. Step S202: Extract key information from the target data to obtain industry dynamic data; Step S203: Update the course content of each training course according to the industry dynamic data.

[0054] In step S201 of some embodiments, the aforementioned target data refers to data used to reflect the latest developments in the industry. Specifically, the target data may be at least one of industry policy data, insurance premium rate adjustment data, and claims-related public opinion data. For example, the aforementioned industry policy data may be policy documents from the China Banking and Insurance Regulatory Commission in various regions, the aforementioned insurance premium rate adjustment data may be insurance premium rate adjustment announcements from insurance companies, and the aforementioned claims-related public opinion data may be claims-related public opinion data from a third-party platform.

[0055] In some embodiments, the aforementioned web crawler refers to a program or script used to automatically crawl web data. The aforementioned crawling of target data via web crawler can involve electronic devices relying on web crawling technology to crawl in real time policy documents from various regional banking and insurance regulatory commissions, insurance company announcements regarding adjustments to insurance rates, and claims-related public opinion data from third-party platforms.

[0056] In step S202 of some embodiments, the aforementioned industry dynamic data refers to data extracted from the target data for updating course content. The aforementioned key information refers to core information in the target data that reflects dynamic changes in the industry. Specifically, the extraction of key information from the target data to obtain industry dynamic data can be achieved by an electronic device using natural language processing technology to extract key information from the target data. For example, key information extracted from policy documents using natural language processing technology could be "updated exemption clauses for new energy vehicle insurance" or "timeliness requirements for claims processing in rainstorm disasters," etc.

[0057] In step S203 of some embodiments, the above-mentioned updating of the course content of each training course according to industry dynamic data can be achieved by the electronic device synchronizing the extracted industry dynamic data to the relevant training courses, so that the course content of each training course is updated as the industry dynamic data changes. For example, after extracting industry dynamic data on "updated exemption clauses for new energy vehicle insurance", the electronic device synchronizes the updated content of the exemption clauses to the "interpretation of new energy vehicle insurance clauses" course, thereby updating the course content of that course.

[0058] In this embodiment, target data is captured in real time by web crawlers and the course content is updated accordingly, ensuring that the course content is synchronized with industry policies, solving the problem of existing training courses being out of touch with actual business, and ensuring the timeliness of the knowledge learned by trainees.

[0059] In one embodiment, before determining the classification label of the training object based on the attribute data of the training object, the method further includes: Obtain the business data and customer profile data of the training subjects; The attribute data of the training subject is determined based on at least one of the business data and the customer profile data.

[0060] In this embodiment, the aforementioned business data refers to data reflecting the business development status of the trainees. Specifically, the business data can be any data that can characterize the trainees' business development capabilities and business direction. For example, the aforementioned business data can be sales data from agents, such as the proportion of insurance policy types, customer policy surrender rate, number of claims assistance, etc.

[0061] In some embodiments, the aforementioned customer profile data refers to data reflecting the characteristics of the customer group served by the training subjects. Specifically, customer profile data can be any data capable of characterizing the distribution and needs of a customer group. For example, the aforementioned customer profile data could include customer age stratification, occupational distribution, insurance preferences, etc.

[0062] In some embodiments, the acquisition of business data and customer profile data of training subjects can be achieved by electronic devices connecting to the insurance company system and core business system via the aforementioned business data channel through an application programming interface (API) to collect business data and customer profile data of training subjects.

[0063] In some embodiments, determining attribute data based on at least one of business data and customer profile data can be achieved by an electronic device establishing a correlation model between agency business capabilities and customer needs, and determining the attribute data of the training subject through this correlation model based on at least one of the business data and customer profile data. For example, the electronic device can determine the attribute data of the training subject as "concentrated in auto insurance business and high proportion of new energy vehicle customers" based on business data and customer profile data showing a high proportion of agency auto insurance policy types and a high proportion of new energy vehicle owners among the customers served.

[0064] In this embodiment, attribute data is determined by business data and customer profile data, so that the subsequently determined classification labels can match the actual business development situation and customer needs of the training subjects, thereby providing accurate input basis for generating targeted training paths.

[0065] In one embodiment, the similarity between the learning data of the training subject and the learning data of the candidate subject is determined in the following way: For each candidate object, a collaborative filtering algorithm is used to calculate the similarity between the learning data of the training object and the learning data of the candidate object.

[0066] In this embodiment, the above-mentioned calculation of the learning data of the training object and the learning data of the candidate object by the collaborative filtering algorithm can be as follows: for each candidate object, the similarity between the learning data of the training object and the learning data of the candidate object is determined based on the first learning completion degree of each course in the course set jointly studied by the training object and the candidate object, and the average learning completion degree of the courses already studied by the training object and the candidate object respectively.

[0067] In this embodiment, the aforementioned set of courses for joint learning refers to a set of courses that both the training subjects and the candidate subjects have studied. The aforementioned first learning completion rate refers to the learning completion rate corresponding to each course in this set of courses. Specifically, the learning completion rate can be a numerical value that can characterize the subject's mastery of the course, such as the percentage of viewing time, a weighted value of test accuracy, etc. The aforementioned average learning completion rate refers to the average learning completion rate of the subject for all the courses they have studied.

[0068] In some embodiments, the determination of the similarity between the training object and the candidate object may be performed by an electronic device using a collaborative filtering algorithm, based on the first learning completion degree and the average learning completion degree.

[0069] Specifically, the similarity between the training subject u and the candidate subject v can be determined by the following formula (1): (1) Where sim(u,v) represents the similarity between the training object u and the candidate object v; uv Let I represent the set of courses that training object u and candidate object v have jointly studied. The above summation operation applies to the course set I. uv Each course in i is conducted; r_vi represents the first learning completion degree of training object u with respect to course i in the course set; r_vi represents the first learning completion degree of candidate object v with respect to course i in the course set. This represents the average learning completion rate of the training subject u across all the courses already studied; This represents the average completion rate of candidate object v across all learned courses.

[0070] For example, for each candidate object, the electronic device calculates the similarity between it and the training object using the above formula (1), and then determines the candidate objects with a similarity higher than 0.7 as similar objects of the training object, thus obtaining a similar agent group.

[0071] In this embodiment, the similarity is determined by the first learning completion rate of the shared learning course and the average learning completion rate of each individual. This allows the similarity to objectively reflect the degree of closeness between the training subjects and the candidate subjects in terms of learning preferences, thereby improving the accuracy of similar subject selection and ensuring the relevance of subsequent training paths.

[0072] In one embodiment, after generating the training path corresponding to the training object, the method further includes at least one of the following: Adjust the order of each training course in the training path according to the learning progress of the trainees; In response to the training subject's test accuracy rate for the first course in the training path being lower than a preset threshold, a supplementary course associated with the first course is added to the training path.

[0073] In this embodiment, the aforementioned learning progress refers to the trainee's completion status of each training course in the training path. Specifically, the learning progress can be any data that reflects the trainee's learning progress, such as the number of completed courses, the currently learned course and its completion rate, etc. The aforementioned adjustment of the order of each training course in the training path according to the learning progress can be achieved by the electronic device dynamically adjusting the order of unlearned training courses in the training path according to the trainee's learning progress.

[0074] In some embodiments, the first course mentioned above refers to any training course in the training path. The preset threshold mentioned above refers to a threshold used to determine whether the test accuracy is too low. The reinforcement course mentioned above refers to a course associated with the first course and used to reinforce the corresponding knowledge points of the first course. Specifically, the reinforcement course can be any course related to the knowledge points of the first course and capable of helping trainees deepen their understanding of those knowledge points.

[0075] In some embodiments, the above-mentioned response to a trainee's test accuracy rate for the first course being lower than a preset threshold, adding a supplementary course associated with the first course to the training path, can be that the electronic device automatically inserts a supplementary course associated with the first course into the training path when it detects that the trainee's test accuracy rate for the first course is lower than the preset threshold. For example, the preset threshold can be 60%. If the trainee's accuracy rate in the "Common Misconceptions in New Energy Vehicle Damage Assessment" micro-course test is lower than 60%, the system will automatically insert a "Damage Assessment Case Breakdown Live Course" as supplementary content.

[0076] In this embodiment, by adjusting the course order according to the learning progress and adding reinforcement courses when the test accuracy is low, the training path is dynamically optimized according to the trainees' learning situation, thereby improving the trainees' mastery of weak knowledge points.

[0077] Please see Figure 4 In some embodiments, after step S104, the method may include, but is not limited to, steps S301 to S305: Step S301: Display a three-dimensional virtual business scene, wherein the three-dimensional virtual business scene includes dynamic interactive elements, and the dynamic interactive elements include at least one of virtual customer status change elements, emergency situation elements, and error reminder elements; Step S302: In response to the training operation performed by the training subject in the three-dimensional virtual business scenario, update the dynamic interactive elements in the three-dimensional virtual business scenario according to the training operation; Step S303: The communication voice between the training subject and the virtual customer in the three-dimensional virtual business scenario is identified by speech recognition to obtain the communication text. Semantic analysis and sentiment analysis are performed on the communication text to generate speech optimization suggestions corresponding to the training operation. Step S304: Obtain training data of the training subject performing the training operation, wherein the training data includes at least one of the communication text, the completion time of the training operation, and the document filling result corresponding to the training operation; Step S305: Evaluate the training effect of the trainees based on the training data.

[0078] In step S301 of some embodiments, the aforementioned three-dimensional virtual business scene refers to a virtual scene obtained by three-dimensionally reproducing a real business scene. Specifically, the three-dimensional virtual business scene can be a virtual scene built based on a 3D engine that reproduces a real business scene. For example, the aforementioned three-dimensional virtual business scene can be a 1:1 business scene model built based on the Unity 3D engine, covering 12 core business scenarios such as 4S store car insurance consultation, on-site investigation of enterprise factory property insurance, communication on claims for burst water pipes in households, and large-scale loss assessment after natural disasters. The aforementioned display of the three-dimensional virtual business scene can be achieved by electronic devices displaying the three-dimensional virtual business scene in an immersive manner through VR devices (such as Pico 4), or by displaying the three-dimensional virtual business scene in a lightweight manner through PCs or mobile devices.

[0079] In some embodiments, the aforementioned dynamic interactive elements refer to elements in a 3D virtual business scenario that can dynamically change and interact with the training subjects. Specifically, dynamic interactive elements include at least one of virtual customer state change elements, unexpected situation elements, and error reminder elements. For example, the aforementioned virtual customer state change element could be a change in customer emotion from skepticism to acceptance; the aforementioned unexpected situation element could be encountering a rainstorm during a survey; and the aforementioned error reminder element could be a document filling error reminder, to simulate complex situations in real business.

[0080] In step S302 of some embodiments, the aforementioned training operation refers to the simulated business operations performed by the trainee in a three-dimensional virtual business scenario. Examples include communicating with virtual customers, filling out claims documents, and inspecting and assessing damaged property. Updating the dynamic interactive elements in the three-dimensional virtual business scenario based on the training operation can be achieved by the electronic device responding to the training operation performed by the trainee by changing the virtual customer's status, triggering unexpected situations, or outputting error alerts. For example, after the trainee accurately explains the deductible clause, the electronic device updates the virtual customer's status from questioning to acceptance; when there are errors in the claims documents filled out by the trainee, the electronic device outputs a document filling error alert, thereby enabling the trainee to complete the training through interaction with the three-dimensional virtual business scenario.

[0081] In step S303 of some embodiments, the above-mentioned generation of speech optimization suggestions corresponding to the training operation can be achieved by embedding an AI speech coaching system based on the Transformer architecture into an electronic device. This system uses speech recognition technology to capture the communication voice between the trainee and the virtual customer in real time and recognize it as communication text. It combines natural language understanding technology to analyze the semantic logic and emotional tendency of the communication text and generates speech optimization suggestions in real time through natural language generation technology. The speech optimization suggestions are then displayed to the trainee through a terminal device.

[0082] In step S304 of some embodiments, the aforementioned training data refers to the data generated by the trainee performing training operations in a three-dimensional virtual business scenario. Specifically, the training data can be any data that can reflect the trainee's practical performance, such as communication text between the trainee and a virtual customer, the completion time of the training operation, and the document completion results corresponding to the training operation.

[0083] In some embodiments, during risk scenario training, electronic devices can also use reinforcement learning algorithms to dynamically adjust the difficulty of the scenario: in the initial stage, a simple scenario is set where the customer only inquires about the basic premium. When the trainee achieves the goal of completing the consultation and facilitating the purchase intention within 10 minutes three times in a row, the system will automatically upgrade the scenario, such as adding complex situations such as the customer questioning the reasonableness of the premium and the timeliness of the claims, or receiving urgent loss assessment requests from other customers during the consultation. Through trial and error, feedback and optimization, the trainee's emergency response capabilities and multi-task coordination capabilities are improved.

[0084] In some embodiments, to resolve the conflict between sharing real-world cases and protecting customer privacy, electronic devices, when accessing a de-identified database of real-world cases, employ federated learning technology to construct a cross-institutional case-sharing alliance. Each insurance company de-identifies its local case data, which may include removing sensitive information such as customer names, ID numbers, and contact information. Through a horizontal federated learning model, case features are shared, allowing trainees to access multi-institutional case training resources without needing to access the original case data. For example, when trainees practice corporate property insurance claims investigation, the system can access cases from different insurance companies, such as factory equipment fire damage assessment and warehouse flood claims, exposing trainees to richer practical scenarios.

[0085] In step S305 of some embodiments, the training effectiveness of the training data can be evaluated by constructing an evaluation hierarchy.

[0086] In this embodiment, by displaying a three-dimensional virtual business scenario containing dynamic interactive elements, the dynamic interactive elements are updated in real time and suggestions for optimizing the scripts are generated during the training operation. The training effect of the trainees is then evaluated based on the training data. The system provides trainees with real-world scenario simulation, claims case analysis, and customer communication script training, which solves the problem of existing training emphasizing theory over practice and having low efficiency in practical application. This helps trainees quickly transform knowledge into business development capabilities.

[0087] In one embodiment, step S305 may include, but is not limited to, steps S401 to S406 as shown below: Step S401: Construct an evaluation hierarchy, wherein the evaluation hierarchy includes a target layer, a criterion layer, and an indicator layer under the criterion layer; Step S402: Perform pairwise comparisons of the evaluation indicators at the same level in the evaluation hierarchy to obtain a judgment matrix; Step S403: Determine the weight of each evaluation index according to the judgment matrix, and perform a consistency check on the judgment matrix; Step S404: Based on the training data, determine the fuzzy relation matrix, wherein the elements in the fuzzy relation matrix are used to characterize the evaluation level of the training object corresponding to the evaluation index; Step S405: If the judgment matrix passes the consistency test, then the weight vector composed of the weights of each evaluation index is combined with the fuzzy relation matrix to perform a fuzzy synthesis operation to obtain the ability evaluation value of the training object. Step S406: Based on the capability assessment value, generate a capability radar chart for the trainee, which is used to characterize the training effect of the trainee.

[0088] In step S401 of some embodiments, the aforementioned evaluation hierarchy refers to a structure obtained by organizing the evaluation indicators hierarchically using the Analytic Hierarchy Process (AHP). Specifically, the evaluation hierarchy includes a target layer, a criterion layer, and an indicator layer under the criterion layer. For example, the target layer can be used to "quantitatively evaluate the effectiveness of agent training," the criterion layer can be used to evaluate primary indicators such as "knowledge mastery," "practical proficiency," "communication skills," and "emergency handling skills," and the indicator layer can be used to evaluate the subdivided indicators under each criterion layer, such as "damage assessment accuracy," "process completion time," and "document completion accuracy" under "practical proficiency," thus forming a clear three-layer structure.

[0089] In step S402 of some embodiments, the above-mentioned comparison of evaluation indicators at the same level to obtain a judgment matrix can be achieved by the electronic device using a 1-9 scale to compare indicators at the same level pairwise. Here, 1 indicates that the two indicators being compared are equally important; 3 indicates that the former is slightly more important than the latter; 5 indicates that the former is significantly more important than the latter; 7 indicates that the former is strongly more important than the latter; 9 indicates that the former is extremely more important than the latter; 2, 4, 6, and 8 are intermediate values; and the reciprocal indicates the opposite comparison, i.e., the latter relative to the former. For example, taking the three secondary indicators A (damage assessment accuracy), B (process completion time), and C (document completion accuracy) under "practical proficiency" as an example, the judgment matrix M obtained by pairwise comparison using the 1-9 scale is as follows: (2) In this judgment matrix, the rows and columns are A, B, and C, respectively, and M_ij represents the importance ratio of indicator i to indicator j. For example, M_12 = 2 indicates that A (loss assessment accuracy) is slightly more important than B (process completion time). The essence of the above judgment matrix is ​​to transform the human perception of "A is slightly more important than B, B is slightly more important than C, and A is significantly more important than C" into a mathematical matrix that can be recognized by a computer.

[0090] The weights of each evaluation index determined by the judgment matrix can be calculated by normalizing the columns of the judgment matrix to determine the largest eigenvalue and its corresponding eigenvector. Specifically, the sum of each column of the judgment matrix can be calculated first, i.e., S1 = 1 + 1 / 2 + 1 / 3 = 1.833, S2 = 2 + 1 + 1 / 2 = 3.5, S3 = 3 + 2 + 1 = 6; then, each element in the judgment matrix M is divided by the sum of its column to obtain the normalized matrix M′ as follows: (3) In this matrix, each element of M′ is a normalized value obtained by dividing the corresponding element in the judgment matrix M by the sum of its columns. Then, each row of the normalized matrix M′ is summed and divided by the number of indicators to obtain an approximate eigenvector X0≈[0.539, 0.297, 0.164]. Here, each element of X0 is an approximate weight of the corresponding indicator, which serves to provide a starting point for subsequent iterative calculations and avoid undirected iteration.

[0091] Then, the judgment matrix M is multiplied by the approximate eigenvector X0, i.e., M·X0, thus transferring the importance preference of the judgment matrix to the vector, resulting in Y0 = [1.625, 0.8945, 0.4922], where each element of Y0 is the result of the operation M·X0. Next, each element of Y0 is divided by the corresponding element of X0, and the average of the resulting ratios is taken to obtain the maximum eigenvalue λmax ≈ 3.009, where λmax is the maximum eigenvalue of the judgment matrix M. It is slightly greater than the number of indicators (3), indicating a slight contradiction in the subjective judgment, but within an acceptable range.

[0092] Furthermore, we need to find the eigenvector corresponding to the largest eigenvalue, that is, solve the equation (M). 3.009I)·X=0, where I is the identity matrix and X is the eigenvector corresponding to the largest eigenvalue. Solving for X, we get X≈[0.542, 0.283, 0.175], which means that in the "practical proficiency" assessment, the original score for the importance of A (damage assessment accuracy) is 0.542, which is more than 3 times that of C (document completion accuracy).

[0093] In steps S403 to S405 of some embodiments, the above-mentioned consistency check of the judgment matrix may be performed by an electronic device calculating the consistency index CI of the judgment matrix and combining it with the average random consistency index RI to calculate the consistency ratio CR, so as to determine whether the consistency is acceptable. The consistency index CI is calculated using the following formula (4): CI = (λmax) n) / (n 1) = (3.009) 3) / (3) 1) = 0.0045 (4) Where CI is the consistency index of the judgment matrix, λmax is the maximum eigenvalue of the judgment matrix, and n is the number of evaluation indicators corresponding to the judgment matrix (n=3 in this example). The smaller the CI value, the more consistent the subjective judgments.

[0094] Then, according to the industry standard, when n=3, the average random consistency index RI=0.58, and the consistency ratio CR is calculated using the following formula (5): CR = CI / RI ≈ 0.0045 / 0.58 ≈ 0.0078 < 0.1 (5) Wherein, CR is the consistency ratio, CI is the consistency index, and RI is the average random consistency index. When CR is less than 0.1, it indicates that the consistency of the judgment matrix is ​​acceptable, that is, the consistency test passes.

[0095] In this embodiment, if the consistency check passes, the electronic device performs a fuzzy synthesis operation on the weight vector composed of the weights of each evaluation index and the fuzzy relation matrix to obtain the competency evaluation value of the training object. Specifically, the competency evaluation value can be calculated using the formula (6) corresponding to the following fuzzy comprehensive evaluation model: (6) Where B is the final comprehensive evaluation result, i.e., the competency assessment value of the trainees; A = [a_1, a_2, ..., a_m] is the weight vector composed of the weights of each evaluation indicator, which is calculated using the above-mentioned analytic hierarchy process. For example, the weight of "damage assessment accuracy" a1 = 0.542, the weight of "process completion time" a2 = 0.283, and the weight of "document completion accuracy" a3 = 0.175; R is a fuzzy relation matrix, where the number of rows m corresponds to the number of secondary indicators under the primary indicator, and the number of columns n corresponds to the number of evaluation levels. For example, if the evaluation levels are divided into four levels: excellent, good, qualified, and unqualified, then n = 4; r ij This represents the membership degree of the i-th secondary indicator at the j-th evaluation level, with a value ranging from 0 to 1. The larger the value, the more the indicator conforms to the characteristics of the corresponding evaluation level (e.g., "loss assessment accuracy = 90%" corresponds to a membership degree r11 = 0.9). The fuzzy synthesis operation is represented in this embodiment, which uses the weighted average method, that is, the elements bj in the evaluation result vector are calculated by the following formula (7): (7) Where bj is the element corresponding to the j-th evaluation level in the evaluation result vector B, ai is the weight of the i-th secondary indicator, and rij is the membership degree of the i-th secondary indicator at the j-th evaluation level. The above summation operation is performed on all m secondary indicators. Through formulas (6) and (7), the electronic device quantifies the subjective performance of the trainees (such as virtual customer satisfaction scores) and objective data (such as practical operation completion time) into a capability evaluation value of 0 to 100 points.

[0096] In step S406 of some embodiments, the above-mentioned generation of the trainee's capability radar chart based on the capability assessment value can be achieved by an electronic device using the capability assessment value corresponding to each criterion layer indicator as the value of each dimension of the radar chart to generate an intuitive personal capability radar chart. For example, the electronic device can use data visualization technology (such as ECharts) to generate the capability radar chart, enabling trainees and institutions to clearly understand their capability shortcomings.

[0097] In some embodiments, the electronic device can also employ data visualization technology to display the training completion rate and distribution of weaknesses of agents in different regions and positions, providing data support for institutions to optimize the allocation of training resources. Furthermore, the electronic device can be equipped with an agent peer review function, using collaborative filtering algorithms to recommend the practical results of agents with similar ability tags, allowing trainees to rate, comment, and exchange information, thus forming a collaborative learning ecosystem.

[0098] In this embodiment, the weights of the evaluation indicators are determined by the analytic hierarchy process and a consistency test is performed. Then, the subjective performance and objective data are quantified into capability assessment values ​​by the fuzzy comprehensive evaluation model, and finally a capability radar chart is generated. This solves the problem that the existing training effect is difficult to quantify and evaluate and that institutions cannot accurately judge the training effectiveness. This makes the training effect traceable and optimizable, and provides data support for the development of personalized reinforcement plans.

[0099] It should be noted that the various embodiments described in this application can be combined with each other or implemented individually without conflict, and this application does not limit this.

[0100] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0101] Please see Figure 6 This application also provides a training device that can implement the above-described training method. The device includes: Module 510 is used to acquire learning data from the trainees. The first determining module 520 is used to determine the similarity between the learning data of the training object and the learning data of each candidate object in a plurality of candidate objects, and to determine the candidate objects whose similarity satisfies the preset similarity conditions as similar objects of the training object. The second determining module 530 is used to determine the classification label of the training object based on the attribute data of the training object; The generation module 540 is used to generate a training path corresponding to the training object based on the course information learned by the similar objects and the classification tags, wherein the training path includes at least one training course. The push module 550 is used to push the training courses to the trainees in the order of the training courses in the training path.

[0102] In this embodiment, the learning data of the trainees is acquired; based on the learning data, similar objects with high similarity to the trainees are identified from multiple candidate objects; simultaneously, classification tags for the trainees are determined based on their attribute data; then, combining the course information learned by similar objects and the classification tags, a personalized training path containing at least one training course is generated for the trainees, and the training courses are pushed to the trainees according to this training path. Thus, by generating training paths for trainees based on the behavior of similar objects and the trainees' own attribute tags, trainees do not need to manually select from a general course library or passively receive all courses. Instead, they directly obtain a combination of courses highly matched to their business scenarios and learning needs, reducing ineffective learning content and repetitive learning time, thereby improving training efficiency.

[0103] The specific implementation of this training device is basically the same as the specific implementation of the training method described above, and will not be repeated here.

[0104] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described training method. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0105] Please see Figure 7 , Figure 7 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes: The processor 601 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application. The memory 602 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 602 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 602 and is called and executed by the processor 601 using the training method of the embodiments of this application. The input / output interface 603 is used to implement information input and output; The communication interface 604 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 605 transmits information between various components of the device (e.g., processor 601, memory 602, input / output interface 603, and communication interface 604); The processor 601, memory 602, input / output interface 603, and communication interface 604 are connected to each other within the device via bus 605.

[0106] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described training method.

[0107] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0108] The training method, training device, electronic device, and storage medium provided in this application acquire the learning data of the training subject; based on the learning data, identify similar objects with high similarity to the training subject from multiple candidate objects; simultaneously, determine the classification label of the training subject based on the attribute data of the training subject; then, combining the course information learned by the similar objects and the classification label, generate a personalized training path for the training subject containing at least one training course, and push the training course to the training subject according to the training path. In this way, by generating a training path for the training subject based on the behavior of similar objects and the training subject's own attribute labels, the training subject does not need to manually select from a general course library or passively receive all courses, but directly obtains a course combination highly matched to their business scenario and learning needs, reducing ineffective learning content and repetitive learning time, thereby improving training efficiency.

[0109] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0110] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0111] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0112] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0113] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0114] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0115] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0116] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0117] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0118] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0119] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A training method, characterized in that, The method includes: Obtain learning data from trainees; Determine the similarity between the learning data of the training object and the learning data of each candidate object among multiple candidate objects, and determine the candidate objects whose similarity satisfies the preset similarity conditions as similar objects of the training object; Based on the attribute data of the training subjects, determine the classification labels of the training subjects; Based on the course information learned by the similar objects and the classification tags, a training path corresponding to the training object is generated, wherein the training path includes at least one training course. The training courses are pushed to the trainees in the order of the training courses in the training path.

2. The method according to claim 1, characterized in that, The method further includes: The target data is obtained by web crawling, including at least one of the following: industry policy data, insurance premium adjustment data, and claims public opinion data. Extract key information from the target data to obtain industry dynamic data; The course content of each training course is updated based on the aforementioned industry dynamic data.

3. The method according to claim 1, characterized in that, The attribute data based on the training subject includes: Obtain the business data and customer profile data of the training subjects; The attribute data of the training subject is determined based on at least one of the business data and the customer profile data.

4. The method according to claim 1, characterized in that, The similarity between the learning data of the training subjects and the learning data of the candidate subjects is determined in the following way: For each candidate object, a collaborative filtering algorithm is used to calculate the similarity between the learning data of the training object and the learning data of the candidate object.

5. The method according to claim 1, characterized in that, After generating the training path corresponding to the training object, the method further includes at least one of the following: Adjust the order of each training course in the training path according to the learning progress of the trainees; In response to the training subject's test accuracy rate for the first course in the training path being lower than a preset threshold, a supplementary course associated with the first course is added to the training path.

6. The method according to claim 1, characterized in that, After generating the training path corresponding to the training object based on the course information learned from the similar objects and the classification tags, the method further includes: Displaying a three-dimensional virtual business scene, wherein the three-dimensional virtual business scene includes dynamic interactive elements, and the dynamic interactive elements include at least one of virtual customer status change elements, emergency situation elements, and error reminder elements; In response to the training operation performed by the training subject in the three-dimensional virtual business scenario, the dynamic interactive elements in the three-dimensional virtual business scenario are updated according to the training operation; The communication voice between the training subject and the virtual customer in the three-dimensional virtual business scenario is identified by speech recognition to obtain the communication text. Semantic analysis and sentiment analysis are performed on the communication text to generate speech optimization suggestions corresponding to the training operation. Obtain training data of the training subject performing the training operation, wherein the training data includes at least one of the communication text, the completion time of the training operation, and the document filling result corresponding to the training operation; Based on the training data, evaluate the training effectiveness of the trainees.

7. The method according to claim 6, characterized in that, The step of evaluating the training effectiveness of the trainees based on the training data includes: Construct an evaluation hierarchy, wherein the evaluation hierarchy includes an objective layer, a criterion layer, and an indicator layer under the criterion layer; The evaluation indicators at the same level in the evaluation hierarchy are compared pairwise to obtain a judgment matrix; The weights of each evaluation index are determined based on the judgment matrix, and the consistency of the judgment matrix is ​​checked. Based on the training data, a fuzzy relation matrix is ​​determined, wherein the elements in the fuzzy relation matrix are used to characterize the evaluation level of the training object corresponding to the evaluation index; If the judgment matrix passes the consistency test, then the weight vector composed of the weights of each evaluation index is combined with the fuzzy relation matrix to perform a fuzzy synthesis operation to obtain the ability evaluation value of the training object. Based on the capability assessment values, a capability radar chart of the trainee is generated, which is used to characterize the training effect of the trainee.

8. A training device, characterized in that, The device includes: The acquisition module is used to acquire the learning data of the trainees; The first determining module is used to determine the similarity between the learning data of the training object and the learning data of each candidate object among multiple candidate objects, and to determine the candidate objects whose similarity satisfies the preset similarity conditions as similar objects of the training object. The second determining module is used to determine the classification label of the training object based on the attribute data of the training object; The generation module is used to generate a training path corresponding to the training object based on the course information learned by the similar objects and the classification tags, wherein the training path includes at least one training course. The push module is used to push the training courses to the trainees in the order of the training courses in the training path.

9. An electronic device, characterized in that, The electronic device includes: a processor and a memory storing computer program instructions; When the processor executes the computer program instructions, it implements the training method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer program instructions that, when executed by a processor, implement the training method as described in any one of claims 1-7.

Images