Insurance claim settlement early warning method, device, equipment and medium
By stitching together interactive and business data during the insurance claims process, performing emotion recognition and anger calculation, and assessing risk levels, this technology addresses the challenges of capturing customer emotional fluctuations and lacking information linkage in existing technologies, thereby achieving efficient claims early warning and improved customer experience.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- PING AN TECH (SHENZHEN) CO LTD
- Filing Date
- 2026-01-09
- Publication Date
- 2026-04-21
AI Technical Summary
Existing insurance claims early warning methods struggle to capture customer emotional fluctuations, lack information linkage, remain at the post-claims stage, and lack automated early warning distribution and closed-loop task management, resulting in low processing efficiency.
By collecting interactive and business data, data splicing, emotion recognition, and anger calculation are performed to assess risk levels, achieving accurate emotion recognition and efficient claims early warning.
It enables accurate identification and timely warning of high-risk customers, improves customer experience and processing efficiency, and ensures customer satisfaction and real-time processing.
Smart Images
Figure CN121901795A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of artificial intelligence technology, and in particular to an insurance claims early warning method, device, equipment, and medium. Background Technology
[0002] Currently, the insurance industry has gradually introduced technologies such as intelligent customer service and voice recognition into its claims services to improve customer experience. However, existing systems still have significant shortcomings: First, most customer service quality inspection and early warning systems rely on keyword matching or simple rule-based judgments, making it difficult to capture subtle emotional fluctuations in customers' voice and text communications, such as implicit dissatisfaction or potential complaint intentions. This leads to delayed early warnings or high false alarm rates. Second, traditional speech-to-text and emotion recognition technologies often operate in isolation, lacking integration with customer profiles and historical interaction records, making it impossible to differentiate risk sensitivities among different customers. Third, current early warning mechanisms typically remain at the post-claims stage, failing to address customer emotions before they escalate, thus impacting complaint mitigation. Furthermore, existing systems lack automated early warning distribution and closed-loop task management, often preventing customer service personnel from receiving effective alerts and taking targeted measures in a timely manner, resulting in decreased customer satisfaction.
[0003] Therefore, existing claims early warning methods are unable to capture emotional fluctuations, lack information linkage, remain at the post-claims stage, and lack automated early warning distribution and closed-loop task management, resulting in low processing efficiency. Summary of the Invention
[0004] This invention provides an insurance claims early warning method, device, equipment, and medium, aiming to solve the problems of low processing efficiency caused by existing claims early warning methods, such as difficulty in capturing emotional fluctuations, lack of information linkage, remaining at the post-claims stage, and lack of automated early warning distribution and task closed-loop management.
[0005] To address the aforementioned problems, in a first aspect, embodiments of the present invention provide an insurance claims early warning method, comprising: Based on the claims warning instructions, the target customers are collected and processed to obtain the corresponding interactive data and business data; The interactive data and the business data are combined to obtain multi-source data. The emotion recognition result is obtained by processing the multi-source data according to the emotion recognition strategy; An anger value is obtained by performing anger calculation processing on the emotion recognition results based on the anger calculation strategy. The risk level is obtained by evaluating the anger level. Claims warnings will be issued to the target customers based on the risk level.
[0006] Secondly, embodiments of this application provide an insurance claim early warning device, which includes: The data collection unit is used to collect and process target customer data based on claims warning instructions to obtain corresponding interactive data and business data; A data splicing unit is used to perform data splicing processing on the interactive data and the business data to obtain multi-source data; The recognition unit is used to process the multi-source data according to the emotion recognition strategy to obtain the emotion recognition result; An anger calculation unit is used to perform anger calculation processing on the emotion recognition result based on the anger calculation strategy to obtain the anger value; An assessment unit is used to perform an assessment process using the anger value to obtain a risk level. The claims early warning unit is used to provide claims early warnings to the target customers based on the risk level.
[0007] Thirdly, embodiments of this application provide a computer device, the computer device including a memory and a processor connected to the memory; the memory is used to store a computer program, and the processor is used to run the computer program stored in the memory to perform the method described in the first aspect above.
[0008] Fourthly, embodiments of this application provide a storage medium storing a computer program, the computer program including program instructions, which, when executed by a processor, implement the method described in the first aspect above.
[0009] This invention provides an insurance claims early warning method, apparatus, device, and medium. The method includes: collecting and processing target customer data based on a claims early warning instruction to obtain corresponding interactive data and business data; performing data splicing processing on the interactive data and business data to obtain multi-source data; performing recognition processing on the multi-source data according to an emotion recognition strategy to obtain an emotion recognition result; performing anger calculation processing on the emotion recognition result based on an anger calculation strategy to obtain an anger value; using the anger value for assessment processing to obtain a risk level; and performing claims early warning processing on the target customer according to the risk level. Therefore, this invention achieves accurate emotion recognition by splicing and processing the collected interactive data and business data to obtain multi-source data, and then performs recognition, anger calculation, and assessment processing on the multi-source data to obtain a risk level. Based on the risk level, claims early warning processing is performed on the target customer. This utilizes multi-source data to achieve accurate emotion recognition, ensuring that high-risk customers are identified and early warnings are distributed immediately, significantly improving customer experience and thus increasing processing efficiency. Attached Figure Description
[0010] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0011] Figure 1 A flowchart illustrating the insurance claims early warning method provided in an embodiment of the present invention; Figure 2 A schematic block diagram of an insurance claim early warning device provided in an embodiment of the present invention; Figure 3 A schematic block diagram of a computer device provided for an embodiment of the present invention. Detailed Implementation
[0012] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0013] It should be understood that, when used in this specification and the appended claims, the terms "comprising" and "including" indicate the presence of the described features, integrals, steps, operations, elements and / or components, but do not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.
[0014] It should also be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.
[0015] It should also be further understood that the term "and / or" as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.
[0016] Please see Figure 1 , Figure 1 This is a flowchart illustrating the insurance claim early warning method provided in an embodiment of the present invention. Figure 1 As shown, this embodiment of the invention provides an insurance claim early warning method, which includes the following steps S110-S160.
[0017] S110. Based on the claims warning instructions, collect and process target customer data to obtain corresponding interactive data and business data.
[0018] In this embodiment, the application scenarios of this solution can be in the financial field, for example, it can be specifically applied to the claims service scenario in the insurance industry.
[0019] In one embodiment, the step of collecting and processing target customer data and obtaining corresponding interaction data and business data based on the claims warning instruction includes: In response to the insurance claim interaction command of the target customer, the claim warning command is generated; The corresponding interactive data and business data are obtained by collecting and processing data from the target customer according to the claim warning instruction.
[0020] In this embodiment, the claim warning instruction is generated in response to the insurance claim interaction instruction of the target customer. Specifically, when the target customer initiates an insurance claim interaction, the insurance claim interaction instruction is automatically generated, and the claim warning instruction is generated based on the insurance claim interaction instruction. The method of initiating the insurance claim interaction can be online customer service dialogue, telephone consultation, etc.
[0021] The process involves collecting and processing data from the target customer according to the claim warning instruction to obtain the corresponding interactive data and business data. Specifically, after receiving the claim warning instruction, the process can collect and process data from the target customer according to the claim warning instruction to obtain the corresponding interactive data and business data. The interactive data may include voice data and dialogue text data. The voice data refers to voice data such as telephone recordings, and the dialogue text data refers to text data such as online customer service dialogue text, emails, and appeal materials. The business data may include data such as the type of insurance claim, the amount of compensation, and the progress milestones.
[0022] As can be seen from the above embodiments, in response to the insurance claim interaction command of the target customer, the claim warning command is generated; and data collection and processing are performed on the target customer according to the claim warning command to obtain the corresponding interaction data and business data. Therefore, real-time response to the insurance claim interaction initiated by the target customer is achieved, ensuring the real-time nature of processing, thereby increasing customer satisfaction and customer stickiness, and improving processing efficiency.
[0023] S120. Perform data splicing processing on the interactive data and the business data to obtain multi-source data.
[0024] In this embodiment, after obtaining the interaction data and the business data, the interaction data and the business data can be spliced together to obtain multi-source data.
[0025] In one embodiment, the step of performing data concatenation processing on the interaction data and the business data to obtain multi-source data includes: The speech data is converted to obtain intermediate speech-text data; The intermediate speech-text data is cleaned to obtain the target speech-text data; The dialogue text data from the historical stages is spliced together to obtain spliced data; The spliced data and the business data are fused together to obtain the target dialogue text data. The target speech text data and the target dialogue text data are used as the multi-source data.
[0026] In this embodiment, the process of converting the speech data to obtain intermediate speech-text data specifically involves using an ASR (Automatic Speech Recognition) engine to transcribe the speech data into intermediate speech-text data while preserving the acoustic features of the speech data. These acoustic features may include pitch, speech rate, volume variations, etc., so that the intermediate speech-text data can be used for subsequent multimodal emotion analysis.
[0027] The intermediate speech text data is cleaned to obtain the target speech text data. Specifically, the intermediate speech text data is processed by word segmentation, noise reduction, stop word filtering, and semantic vectorization to complete the cleaning process and obtain the target speech text data, ensuring the accuracy of the data.
[0028] The process of splicing the dialogue text data from historical stages to obtain spliced data involves acquiring the conversation history and splicing together the interactive dialogues of the target customer in multiple stages such as reporting the incident, checking the claim progress, and confirming the payment to obtain the spliced data. This forms a complete context, avoids misjudgment based on single fragments, ensures the accuracy of subsequent processing, and thus improves processing efficiency.
[0029] The process of fusing the spliced data and the business data to obtain target dialogue text data involves aligning the spliced data with the structured business data (such as case type, compensation amount, and progress node) to obtain the target dialogue text data, which forms an input for subsequent emotion and profile analysis.
[0030] The target speech text data and the target dialogue text data are used as the multi-source data, and subsequent targeted processing is performed based on the multi-source data.
[0031] Through the above embodiments, it can be seen that the voice data is converted to obtain intermediate voice-text data; the intermediate voice-text data is cleaned to obtain target voice-text data; the dialogue text data from historical stages is spliced to obtain spliced data; the spliced data and the business data are fused to obtain target dialogue text data; and the target voice-text data and the target dialogue text data are used as the multi-source data. Therefore, the conversion, cleaning, splicing, and fusion of interactive data and business data are achieved, ensuring data accuracy, increasing customer satisfaction and customer loyalty, and thus improving processing efficiency.
[0032] S130. The multi-source data is processed according to the emotion recognition strategy to obtain the emotion recognition result.
[0033] In this embodiment, after obtaining the multi-source data, the multi-source data can be processed according to an emotion recognition strategy to obtain an emotion recognition result.
[0034] In one embodiment, the step of processing the multi-source data according to an emotion recognition strategy to obtain an emotion recognition result includes: The target dialogue text data is semantically processed using a pre-trained semantic recognition model to obtain the text sentiment. The target speech text data is processed using a pre-trained acoustic recognition model to obtain the voice emotion. The emotional distribution vector is obtained by performing an emotional fusion process on the text emotion and the voice emotion. The emotion recognition result is obtained by analyzing and processing the emotion distribution vector.
[0035] In this embodiment, the text sentiment is obtained by performing semantic recognition processing on the target dialogue text data using a pre-trained semantic recognition model. Specifically, the text sentiment is obtained by performing semantic recognition processing on the target dialogue text data using the semantic recognition model. The pre-trained semantic recognition model is a fine-tuned LLM (Large Language Model) specific to the financial field, possessing the ability to understand contextual semantics and identify risk intent. The text sentiment can include objective complaints or subjective attacks; customer complaints can be such as slow claims processing, while subjective attacks can be such as strong accusations.
[0036] The process involves using a pre-trained acoustic recognition model to perform emotion recognition processing on the target speech text data to obtain the voice emotion. Specifically, the pre-trained acoustic recognition model can be a model obtained by modeling changes in volume, pitch, and speech rate using an emotion recognition network (BiLSTM + Attention). The voice emotion can include subdivided emotions such as anger, anxiety, and helplessness.
[0037] The process of fusing the text emotion and the voice emotion to obtain an emotion distribution vector involves, specifically, using a Late Fusion strategy, weighted fusing the text emotion and the voice emotion output to obtain the real-time emotion distribution vector of the target customer.
[0038] The emotion recognition result is obtained by analyzing and processing the emotion distribution vector. Specifically, the evolution trend of the emotion recognition result in the entire claims process is analyzed according to the time series to obtain the emotion recognition result. The emotion recognition result may include initial calmness, mid-term anxiety, and late-term anger.
[0039] As demonstrated by the above embodiments, text emotion is obtained by semantic recognition processing of the target dialogue text data using a pre-trained semantic recognition model; voice emotion is obtained by emotion recognition processing of the target speech text data using a pre-trained acoustic recognition model; the text emotion and the voice emotion are fused to obtain an emotion distribution vector; and the emotion distribution vector is analyzed to obtain the emotion recognition result. Therefore, by accurately obtaining text emotion and voice emotion through semantic and acoustic recognition models of the target dialogue text data and target speech text data, precise emotion recognition is achieved, ensuring processing accuracy and thus improving processing efficiency.
[0040] S140. Based on the anger calculation strategy, the emotion recognition result is processed to obtain the anger value.
[0041] In this embodiment, after determining the emotion recognition result, the emotion recognition result can be processed to obtain an anger value based on an anger calculation strategy.
[0042] In one embodiment, the anger calculation process performed on the emotion recognition result based on the anger calculation strategy to obtain an anger value includes: Obtain the claims data of the target customer, and perform probability calculation on the claims data based on a probability calculation strategy to obtain a sensitivity index; The weight of the current interaction scenario is determined based on the interaction data; An emotion intensity score is obtained by scoring the emotion recognition results. The anger value is obtained by calculating the anger value using the anger value calculation formula, which is applied to the sensitivity index, the weight of the current interaction scenario, and the emotion intensity score.
[0043] In this embodiment, the acquisition of the target customer's claims data and the processing of the claims data using a probability calculation strategy to obtain a sensitivity index are specifically achieved by acquiring and processing the target customer's dialogue text data, and then performing probability calculations on the claims data to obtain the target customer's sensitivity index. The claims data includes basic attributes, historical claims behavior, and financial relationships. Basic attributes include age and occupation; historical claims behavior includes whether there have been multiple claims and whether there have been complaint records; and financial relationships include loan / policy value.
[0044] The step of determining the weight of the current interaction scenario based on the interaction data specifically involves obtaining the weight of the current interaction scenario based on the interaction data. The weight of the current interaction scenario may include the weight of the compensation dispute scenario, the weight of the progress query scenario, etc. Corresponding weight values can be preset for the weight of the compensation dispute scenario and the weight of the progress query scenario. For example, the weight of the compensation dispute scenario is greater than the weight of the progress query scenario.
[0045] The process of scoring based on the emotion recognition results to obtain an emotion intensity score involves, specifically, adding the scores of multiple emotions from the emotion recognition results together to obtain the emotion intensity score. An emotion value can be pre-set for each emotion. The emotion recognition results include multiple emotions, such as peace, anxiety, anger, and helplessness.
[0046] The anger value is obtained by calculating the anger value using the anger value calculation formula, which applies the sensitivity index, the current interaction scenario weight, and the emotion intensity score. Specifically, the anger value calculation formula is as follows: Where A is the anger value; α, β, and γ are coefficients adjusted according to actual business conditions; E is the emotion intensity score; P is the sensitivity index; and S is the weight of the current interaction scenario. The anger value is obtained by calculating the anger value using the anger value calculation formula, which is applied to the sensitivity index, the weight of the current interaction scenario, and the emotion intensity score.
[0047] Through the above embodiments, it can be seen that: The claims data of the target customer is acquired, and a sensitivity index is obtained by performing probability calculation on the claims data based on a probability calculation strategy; the weight of the current interaction scenario is determined based on the interaction data; an emotion intensity score is obtained by scoring the emotion recognition results; and an anger value is obtained by calculating the anger value using the anger value calculation formula on the sensitivity index, the weight of the current interaction scenario, and the emotion intensity score. Therefore, an accurate anger value is obtained by calculating the anger value through the sensitivity index, the weight of the current interaction scenario, and the emotion intensity score, realizing the quantification of anger value and providing a quantitative indicator for subsequent evaluation and processing, thereby ensuring the accuracy of subsequent processing and improving processing efficiency.
[0048] S150. The risk level is obtained by evaluating the anger value.
[0049] In this embodiment, after obtaining the anger value, the anger value can be used for evaluation to obtain the risk level.
[0050] In one embodiment, the process of using the anger value to assess and obtain a risk level includes: Get the preset rage threshold set; The risk level is obtained by comparing the anger value with the preset anger threshold set; wherein the risk level includes low risk, medium risk and high risk.
[0051] In this embodiment, a preset anger threshold set is obtained, wherein the preset anger threshold set may include a first anger threshold range, a second anger threshold range, and a third anger threshold range; the anger value is compared with the preset anger threshold set to obtain the risk level. Specifically, if the anger value is within the first anger threshold range, the risk level is low risk; if the anger value is within the second anger threshold range, the risk level is medium risk; if the anger value is within the third anger threshold range, the risk level is high risk; wherein the first anger threshold range is smaller than the second anger threshold range, and the second anger threshold range is smaller than the third anger threshold range; the risk level includes low risk, medium risk, and high risk.
[0052] As illustrated in the above embodiments, a preset anger threshold set is obtained; the risk level is obtained by comparing the anger value with the preset anger threshold set; wherein the risk level includes low risk, medium risk, and high risk. Therefore, by obtaining an accurate risk level through the comparison of the anger value with the preset anger threshold set, precise insurance claim warnings for different risk levels can be achieved. Closed-loop distribution ensures implementation and forms a complete intelligent consumer protection warning ecosystem, thereby ensuring processing efficiency.
[0053] S160. Based on the risk level, conduct claim warning processing for the target customer.
[0054] In this embodiment, after obtaining the risk level, a claims warning can be issued to the target customer based on the risk level.
[0055] In one embodiment, the step of providing a claim warning to the target customer based on the risk level includes: If the risk level is low, then the target customer will be given a claim warning according to the first processing strategy; If the risk level is medium risk, then the target customer will be given a claim warning according to the second processing strategy; If the risk level is high, then the target customer will be given a claim warning according to the third processing strategy.
[0056] In this embodiment, when the risk level is low, a claims warning is issued to the target customer according to the first processing strategy; wherein, the first processing strategy may be routine reassurance, etc. When the risk level is medium, a claims warning is issued to the target customer according to the second processing strategy; wherein, the second processing strategy may be manual monitoring of progress, etc. When the risk level is high, a claims warning is issued to the target customer according to the third processing strategy; wherein, the third processing strategy may be immediate distribution of the warning to the specific executor and direct communication with the target customer, etc.
[0057] As can be seen from the above embodiments, if the risk level is low, a claim warning is issued to the target customer according to the first processing strategy; if the risk level is medium, a claim warning is issued to the target customer according to the second processing strategy; and if the risk level is high, a claim warning is issued to the target customer according to the third processing strategy. Therefore, different processing strategies for different risk levels are implemented to proactively trigger claim warnings, improve response speed and customer satisfaction, and thus ensure processing efficiency.
[0058] Furthermore, based on the case type and risk level of the target customer, the system can automatically match the customer to the responsible department (such as front-line customer service supervisor, claims specialist, or consumer protection compliance department) and push the information in real time through the company's internal messaging system to ensure processing efficiency.
[0059] After processing the claim warning for the target customer based on the risk level, the method further includes: retrieving a corresponding target script template from a preset reassurance script template library according to the first processing strategy, the second processing strategy, or the third processing strategy, to provide personalized response suggestions to specific personnel. The target script template may include templates such as progress explanation templates and apology / compensation templates.
[0060] After issuing claim warnings to the target customers based on the risk level, the process further includes: recording each warning response to obtain a warning log, which may include response time, effectiveness of measures, changes in customer sentiment, etc., and optimizing subsequent warning thresholds and distribution strategies through a feedback learning mechanism. Simultaneously, if the sentiment of high-risk customers does not subside, the legal compliance team can be triggered in advance to achieve "preventive mediation" and reduce litigation risks.
[0061] In summary, this embodiment of the invention collects and processes corresponding interactive data and business data from target customers based on claims warning instructions; it then performs data splicing on the interactive data and business data to obtain multi-source data; it performs recognition processing on the multi-source data according to an emotion recognition strategy to obtain an emotion recognition result; it performs anger calculation processing on the emotion recognition result based on an anger calculation strategy to obtain an anger value; it uses the anger value for assessment processing to obtain a risk level; and it performs claims warning processing on the target customer according to the risk level. Therefore, this embodiment of the invention obtains multi-source data by splicing the collected interactive data and business data, performs recognition, anger calculation, and assessment processing on the multi-source data to obtain a risk level, and performs claims warning processing on the target customer according to the risk level. This utilizes multi-source data to achieve accurate emotion recognition, ensuring that high-risk customers are identified and warned in a timely manner, significantly improving customer experience and thus increasing processing efficiency.
[0062] Figure 2 This is a schematic block diagram of an insurance claim early warning device provided in an embodiment of the present invention. Figure 2 As shown, this embodiment of the invention provides an insurance claim early warning device 700 that implements the method described above. Specifically, please refer to... Figure 2 The insurance claim early warning device 700 includes: The data collection unit 701 is used to collect and process target customer data based on claims warning instructions to obtain corresponding interactive data and business data; The data splicing unit 702 is used to perform data splicing processing on the interactive data and the business data to obtain multi-source data; The recognition unit 703 is used to process the multi-source data according to the emotion recognition strategy to obtain the emotion recognition result; Anger calculation unit 704 is used to perform anger calculation processing on the emotion recognition result based on an anger calculation strategy to obtain an anger value. The assessment unit 705 is used to perform an assessment process using the anger value to obtain a risk level; The claims warning unit 706 is used to provide claims warnings to the target customer based on the risk level.
[0063] In some embodiments, when the data splicing unit 702 performs the step of splicing the interactive data and the business data to obtain multi-source data, it is specifically used for: The speech data is converted to obtain intermediate speech-text data; The intermediate speech-text data is cleaned to obtain the target speech-text data; The dialogue text data from the historical stages is spliced together to obtain spliced data; The spliced data and the business data are fused together to obtain the target dialogue text data. The target speech text data and the target dialogue text data are used as the multi-source data.
[0064] In some embodiments, when the recognition unit 703 performs the processing step of recognizing the multi-source data according to the emotion recognition strategy to obtain the emotion recognition result, it is specifically used for: The target dialogue text data is semantically processed using a pre-trained semantic recognition model to obtain the text sentiment. The target speech text data is processed using a pre-trained acoustic recognition model to obtain the voice emotion. The emotional distribution vector is obtained by performing an emotional fusion process on the text emotion and the voice emotion. The emotion recognition result is obtained by analyzing and processing the emotion distribution vector.
[0065] In some embodiments, when the anger calculation unit 704 performs the step of calculating the anger value based on the emotion recognition result using an anger calculation strategy to obtain an anger value, it is specifically used for: Obtain the claims data of the target customer, and perform probability calculation on the claims data based on a probability calculation strategy to obtain a sensitivity index; The weight of the current interaction scenario is determined based on the interaction data; An emotion intensity score is obtained by scoring the emotion recognition results. The anger value is obtained by calculating the anger value using the anger value calculation formula, which is applied to the sensitivity index, the weight of the current interaction scenario, and the emotion intensity score.
[0066] In some embodiments, when performing the step of obtaining a risk level by evaluating the anger value, the assessment unit 705 is specifically used for: Get the preset rage threshold set; The risk level is obtained by comparing the anger value with the preset anger threshold set; wherein the risk level includes low risk, medium risk and high risk.
[0067] In some embodiments, when the claims warning unit 706 performs the processing step of issuing a claims warning to the target customer based on the risk level, it is specifically used for: If the risk level is low, then the target customer will be given a claim warning according to the first processing strategy; If the risk level is medium risk, then the target customer will be given a claim warning according to the second processing strategy; If the risk level is high, then the target customer will be given a claim warning according to the third processing strategy.
[0068] In some embodiments, when the collection unit 701 performs the processing step of collecting and processing target customer data and corresponding interactive data based on the claims warning instruction to obtain corresponding interactive data and business data, it is specifically used for: In response to the insurance claim interaction command of the target customer, the claim warning command is generated; The corresponding interactive data and business data are obtained by collecting and processing data from the target customer according to the claim warning instruction.
[0069] It should be noted that those skilled in the art can clearly understand that the specific implementation process of the above-mentioned device can be referred to the corresponding description in the foregoing method embodiments. For the sake of convenience and brevity, it will not be repeated here.
[0070] The above-described device can be implemented as a computer program, and the computer program can be implemented in, for example... Figure 3 It runs on the computer device shown.
[0071] Please see Figure 3 , Figure 3 This is a schematic block diagram of an electronic device provided in an embodiment of the present invention. The electronic device 800 can be a terminal or a server. The terminal can be an electronic device with communication functions. The server can be a standalone server or a server cluster composed of multiple servers.
[0072] See Figure 3The electronic device 800 includes a processor 802, a memory, and a network interface 805 connected via a system bus 801. The memory may include a non-volatile storage medium 803 and internal memory 804.
[0073] The non-volatile storage medium 803 may store an operating system 8031 and a computer program 8032. The computer program 8032 includes program instructions that, when executed, cause the processor 802 to perform an insurance claim early warning method.
[0074] The processor 802 provides computing and control capabilities to support the operation of the entire electronic device 800.
[0075] The internal memory 804 provides an environment for the operation of the computer program 8032 in the non-volatile storage medium 803. When the computer program 8032 is executed by the processor 802, the processor 802 can execute an insurance claim early warning method.
[0076] This network interface 805 is used for network communication with other devices. Those skilled in the art will understand that... Figure 3 The structure shown is merely a block diagram of a portion of the structure related to the present invention and does not constitute a limitation on the electronic device 800 to which the present invention is applied. The specific electronic device 800 may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.
[0077] The processor 802 is used to run a computer program 8032 stored in the memory to perform the following steps: Based on the claims warning instructions, the target customers are collected and processed to obtain the corresponding interactive data and business data; The interactive data and the business data are combined to obtain multi-source data. The emotion recognition result is obtained by processing the multi-source data according to the emotion recognition strategy; An anger value is obtained by performing anger calculation processing on the emotion recognition results based on the anger calculation strategy. The risk level is obtained by evaluating the anger level. Claims warnings will be issued to the target customers based on the risk level.
[0078] In some embodiments, when implementing the step of concatenating the interactive data and the business data to obtain multi-source data, the processor 802 is specifically used for: The speech data is converted to obtain intermediate speech-text data; The intermediate speech-text data is cleaned to obtain the target speech-text data; The dialogue text data from the historical stages is spliced together to obtain spliced data; The spliced data and the business data are fused together to obtain the target dialogue text data. The target speech text data and the target dialogue text data are used as the multi-source data.
[0079] In some embodiments, when implementing the processing step of processing the multi-source data according to the emotion recognition strategy to obtain the emotion recognition result, the processor 802 is specifically used for: The target dialogue text data is semantically processed using a pre-trained semantic recognition model to obtain the text sentiment. The target speech text data is processed using a pre-trained acoustic recognition model to obtain the voice emotion. The emotional distribution vector is obtained by performing an emotional fusion process on the text emotion and the voice emotion. The emotion recognition result is obtained by analyzing and processing the emotion distribution vector.
[0080] In some embodiments, when implementing the step of calculating the anger value by performing anger calculation on the emotion recognition result using the anger calculation strategy, the processor 802 is specifically used to: Obtain the claims data of the target customer, and perform probability calculation on the claims data based on a probability calculation strategy to obtain a sensitivity index; The weight of the current interaction scenario is determined based on the interaction data; An emotion intensity score is obtained by scoring the emotion recognition results. The anger value is obtained by calculating the anger value using the anger value calculation formula, which is applied to the sensitivity index, the weight of the current interaction scenario, and the emotion intensity score.
[0081] In some embodiments, when implementing the step of obtaining a risk level by evaluating the anger value, the processor 802 is specifically configured to: Get the preset rage threshold set; The risk level is obtained by comparing the anger value with the preset anger threshold set; wherein the risk level includes low risk, medium risk and high risk.
[0082] In some embodiments, when implementing the processing step of providing a claims warning to the target customer based on the risk level, the processor 802 is specifically used for: If the risk level is low, then the target customer will be given a claim warning according to the first processing strategy; If the risk level is medium risk, then the target customer will be given a claim warning according to the second processing strategy; If the risk level is high, then the target customer will be given a claim warning according to the third processing strategy.
[0083] In some embodiments, when implementing the processing step of collecting and processing target customer data and corresponding interactive data based on the claims warning instruction, the processor 802 is specifically used for: In response to the insurance claim interaction command of the target customer, the claim warning command is generated; The corresponding interactive data and business data are obtained by collecting and processing data from the target customer according to the claim warning instruction.
[0084] It should be understood that, in this embodiment of the invention, the processor 802 may be a Central Processing Unit (CPU), or it may be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor.
[0085] It will be understood by those skilled in the art that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program includes program instructions and can be stored in a storage medium, which is a computer-readable storage medium. The program instructions are executed by at least one processor in the computer system to implement the process steps of the embodiments of the above methods.
[0086] Therefore, the present invention also provides a storage medium. This storage medium can be a computer-readable storage medium. The storage medium stores a computer program, wherein the computer program includes program instructions. When executed by a processor, the program instructions cause the processor to perform the following steps: Based on the claims warning instructions, the target customers are collected and processed to obtain the corresponding interactive data and business data; The interactive data and the business data are combined to obtain multi-source data. The emotion recognition result is obtained by processing the multi-source data according to the emotion recognition strategy; An anger value is obtained by performing anger calculation processing on the emotion recognition results based on the anger calculation strategy. The risk level is obtained by evaluating the anger level. Claims warnings will be issued to the target customers based on the risk level.
[0087] In one embodiment, when the processor executes the program instructions to implement the processing step of concatenating the interactive data and the business data to obtain multi-source data, it is specifically used for: The speech data is converted to obtain intermediate speech-text data; The intermediate speech-text data is cleaned to obtain the target speech-text data; The dialogue text data from the historical stages is spliced together to obtain spliced data; The spliced data and the business data are fused together to obtain the target dialogue text data. The target speech text data and the target dialogue text data are used as the multi-source data.
[0088] In one embodiment, when the processor executes the program instructions to implement the processing step of identifying and processing the multi-source data according to the emotion recognition strategy to obtain the emotion recognition result, it is specifically used for: The target dialogue text data is semantically processed using a pre-trained semantic recognition model to obtain the text sentiment. The target speech text data is processed using a pre-trained acoustic recognition model to obtain the voice emotion. The emotional distribution vector is obtained by performing an emotional fusion process on the text emotion and the voice emotion. The emotion recognition result is obtained by analyzing and processing the emotion distribution vector.
[0089] In one embodiment, when the processor executes the program instructions to implement the step of calculating the anger value based on the anger calculation strategy on the emotion recognition result, it is specifically used for: Obtain the claims data of the target customer, and perform probability calculation on the claims data based on a probability calculation strategy to obtain a sensitivity index; The weight of the current interaction scenario is determined based on the interaction data; An emotion intensity score is obtained by scoring the emotion recognition results. The anger value is obtained by calculating the anger value using the anger value calculation formula, which is applied to the sensitivity index, the weight of the current interaction scenario, and the emotion intensity score.
[0090] In one embodiment, when the processor executes the program instructions to implement the step of evaluating the risk level using the anger value, it is specifically used for: Get the preset rage threshold set; The risk level is obtained by comparing the anger value with the preset anger threshold set; wherein the risk level includes low risk, medium risk and high risk.
[0091] In one embodiment, when the processor executes the program instructions to implement the processing step of providing a claim warning to the target customer based on the risk level, it is specifically used for: If the risk level is low, then the target customer will be given a claim warning according to the first processing strategy; If the risk level is medium risk, then the target customer will be given a claim warning according to the second processing strategy; If the risk level is high, then the target customer will be given a claim warning according to the third processing strategy.
[0092] In one embodiment, when the processor executes the program instructions to implement the processing step of collecting and processing target customer data and obtaining corresponding interactive data and business data based on the claims warning instruction, it is specifically used for: In response to the insurance claim interaction command of the target customer, the claim warning command is generated; The corresponding interactive data and business data are obtained by collecting and processing data from the target customer according to the claim warning instruction.
[0093] The storage medium can be any computer-readable storage medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory (ROM), magnetic disk, or optical disk.
[0094] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.
[0095] In the several embodiments provided by this invention, 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 example, the division of each unit is merely a logical functional division, and there may be other division methods in actual implementation. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed.
[0096] The steps in the method of this invention can be adjusted, merged, or reduced in order according to actual needs. The units in the device of this invention can be merged, divided, or reduced according to actual needs. Furthermore, the functional units in the various embodiments of this invention 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.
[0097] 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 storage medium. Based on this understanding, the technical solution of the present invention, 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 several instructions to cause an electronic device (which may be a personal computer, a terminal, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention.
[0098] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims. The software tools, models, or components appearing in the embodiments of the present invention are merely illustrative examples and do not represent actual use.
Claims
1. An insurance claims early warning method, characterized in that, include: Based on the claims warning instructions, the target customers are collected and processed to obtain the corresponding interactive data and business data; The interactive data and the business data are combined to obtain multi-source data. The emotion recognition result is obtained by processing the multi-source data according to the emotion recognition strategy; An anger value is obtained by performing anger calculation processing on the emotion recognition results based on the anger calculation strategy. The risk level is obtained by evaluating the anger level. Claims warnings will be issued to the target customers based on the risk level.
2. The method according to claim 1, characterized in that, The interactive data includes voice data and dialogue text data. The process of concatenating the interactive data and the business data to obtain multi-source data includes: The speech data is converted to obtain intermediate speech-text data; The intermediate speech-text data is cleaned to obtain the target speech-text data; The dialogue text data from the historical stages is spliced together to obtain spliced data; The spliced data and the business data are fused together to obtain the target dialogue text data. The target speech text data and the target dialogue text data are used as the multi-source data.
3. The method according to claim 2, characterized in that, The step of processing the multi-source data according to the emotion recognition strategy to obtain the emotion recognition result includes: The target dialogue text data is semantically processed using a pre-trained semantic recognition model to obtain the text sentiment. The target speech text data is processed using a pre-trained acoustic recognition model to obtain the voice emotion. The emotional distribution vector is obtained by performing an emotional fusion process on the text emotion and the voice emotion. The emotion recognition result is obtained by analyzing and processing the emotion distribution vector.
4. The method according to claim 1, characterized in that, The anger calculation process, based on the anger calculation strategy, performs anger calculation on the emotion recognition result to obtain an anger value, including: Obtain the claims data of the target customer, and perform probability calculation on the claims data based on a probability calculation strategy to obtain a sensitivity index; The weight of the current interaction scenario is determined based on the interaction data; An emotion intensity score is obtained by scoring the emotion recognition results. The anger value is obtained by calculating the anger value using the anger value calculation formula, which is applied to the sensitivity index, the weight of the current interaction scenario, and the emotion intensity score.
5. The method according to claim 1, characterized in that, The risk level is obtained by assessing the anger value, including: Get the preset rage threshold set; The risk level is obtained by comparing the anger value with the preset anger threshold set; wherein the risk level includes low risk, medium risk and high risk.
6. The method according to claim 5, characterized in that, The process of issuing early warnings for claims against the target customer based on the risk level includes: If the risk level is low, then the target customer will be given a claim warning according to the first processing strategy; If the risk level is medium risk, then the target customer will be given a claim warning according to the second processing strategy; If the risk level is high, then the target customer will be given a claim warning according to the third processing strategy.
7. The method according to claim 1, characterized in that, The process of collecting and processing target customer data based on claims warning instructions to obtain corresponding interactive data and business data includes: In response to the insurance claim interaction command of the target customer, the claim warning command is generated; The corresponding interactive data and business data are obtained by collecting and processing data from the target customer according to the claim warning instruction.
8. An insurance claim early warning device, characterized in that, include: The data collection unit is used to collect and process target customer data based on claims warning instructions to obtain corresponding interactive data and business data; A data splicing unit is used to perform data splicing processing on the interactive data and the business data to obtain multi-source data; The recognition unit is used to process the multi-source data according to the emotion recognition strategy to obtain the emotion recognition result; An anger calculation unit is used to perform anger calculation processing on the emotion recognition result based on the anger calculation strategy to obtain the anger value; An assessment unit is used to evaluate the anger value to obtain a risk level. The claims early warning unit is used to provide claims early warnings to the target customers based on the risk level.
9. A computer device, characterized in that, The computer device includes a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the method as described in any one of claims 1-7.
10. A storage medium, characterized in that, The storage medium stores a computer program, which includes program instructions that, when executed by a processor, can implement the method as described in any one of claims 1-7.