Automobile protection delay service cost optimization method and system based on big data analysis

Abstract

The invention discloses an automobile insurance delay service cost optimization method and system based on big data analysis, and particularly relates to the technical field of cost optimization. The method comprises the following steps of: performing semantic analysis on historical maintenance record data of a vehicle to generate a fault feature vector; performing time sequence alignment processing on the real-time vehicle-mounted diagnosis data stream, and constructing a time sequence fault feature; dividing a high-frequency fault mode and a low-frequency fault mode based on the fault feature vector and the time sequence fault feature, and respectively constructing a high-frequency fault probability distribution model and a low-frequency fault risk factor set; for the high-frequency fault mode, calculating expected maintenance cost by adopting a Monte Carlo simulation method; for the low-frequency fault mode, constructing a fault propagation network based on topological correlation analysis; based on the expected maintenance cost and the fault propagation network, an extended insurance cost decision matrix is output through risk loss function optimization; an insurance delay service pricing strategy is generated in combination with an insurance actuarial rule base, and the accuracy and stability of insurance delay cost evaluation are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of cost optimization, and more specifically, to a method and system for optimizing the cost of vehicle extended warranty services based on big data analysis. Background Art

[0002] In the context of the rapid development of vehicle extended warranty services, extended warranty cost control and risk pricing have become the core of industry operations. Existing extended warranty analysis methods mainly rely on statistical modeling of high-frequency common faults, and it is difficult to effectively model and evaluate fault events with low occurrence frequency but high compensation amounts, resulting in the lack of stability and rationality of extended warranty pricing strategies and affecting the user protection experience. Summary of the Invention

[0003] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a method and system for optimizing the cost of vehicle extended warranty services based on big data analysis to solve the problems raised in the above background art.

[0004] To achieve the above object, the present invention provides the following technical solutions: A method for optimizing the cost of vehicle extended warranty services based on big data analysis, comprising the following steps: S1: Perform semantic parsing on vehicle historical maintenance record data to generate a fault feature vector; perform time series alignment processing on real-time on-vehicle diagnostic data streams and output time series fault features; S2: Based on the fault feature vector and time series fault features, divide high-frequency fault modes and low-frequency fault modes, and generate a high-frequency fault probability distribution model and a low-frequency fault risk factor set; S3: According to the high-frequency fault probability distribution model, calculate the expected maintenance cost in the high-frequency fault scenario through Monte Carlo simulation; S4: For the low-frequency fault risk factor set, perform topological correlation analysis on sparse fault events, identify component-level fault chain reaction paths, and generate a fault propagation network for low-frequency high-cost events; S5: Based on the expected maintenance cost and the fault propagation network, optimize and output an extended warranty cost decision matrix through a risk loss function; S6: According to the extended warranty cost decision matrix, call a preset insurance actuarial rule library to generate an extended warranty service pricing strategy.

[0005] In a preferred embodiment, S1 is specifically: Collect vehicle historical maintenance record data; Use natural language processing algorithms to perform semantic parsing and feature extraction on vehicle historical maintenance record data to generate a fault feature vector; Collect real-time on-vehicle diagnostic data streams; Perform time series alignment processing on real-time in-vehicle diagnostic data streams based on time windows to generate vehicle timing fault features.

[0006] In a preferred embodiment, S2 is specifically as follows: Analyze the fault feature vectors and vehicle timing fault features through a deep neural network algorithm to generate a classification result of the fault modes of automotive components; Set a fault frequency threshold based on the classification result of the fault modes of automotive components, and classify the fault modes of automotive components into high-frequency fault modes and low-frequency fault modes based on the fault frequency threshold; Based on the high-frequency fault modes, construct a high-frequency fault probability distribution model through probability statistical methods; Based on the low-frequency fault modes, extract low-frequency fault risk features through statistical analysis methods to generate a set of low-frequency fault risk factors.

[0007] In a preferred embodiment, S3 is specifically as follows: Establish a stochastic simulation model of high-frequency fault scenarios of automotive components based on the Monte Carlo simulation method; Perform multiple iterative runs on the stochastic simulation model of high-frequency fault scenarios of automotive components according to a preset number of simulation times; During each iterative run, randomly generate fault occurrence events that conform to the high-frequency fault probability distribution model, and determine the single repair cost of each fault event; Statistically analyze the single repair costs of all iterative runs to calculate the expected repair cost of high-frequency fault scenarios of automotive components.

[0008] In a preferred embodiment, S4 is specifically as follows: Based on the set of low-frequency fault risk factors, construct a topological association model of low-frequency faults of automotive components using a spatial reinforcement topology algorithm; Analyze the mutual association relationships between low-frequency faults of automotive components using the topological association model of low-frequency faults of automotive components to determine the association strength between low-frequency faults of automotive components; Identify the chain reaction paths of low-frequency faults of automotive components according to the association strength to generate a fault propagation network of low-frequency high-cost fault events of automotive components.

[0009] In a preferred embodiment, S5 is specifically as follows: Based on the expected repair cost of high-frequency fault scenarios of automotive components and the fault propagation network of low-frequency high-cost fault events of automotive components, construct a risk loss function; Use the risk loss function to perform optimization analysis on the risk weight coefficients of high-frequency fault scenarios of automotive components and low-frequency high-cost fault events of automotive components; Calculate the expected value of the extended warranty cost corresponding to the high-frequency failure scenarios of automotive parts and the low-frequency and high-cost failure events of automotive parts according to the risk weight coefficient obtained from the optimization analysis; Generate an extended warranty cost decision matrix for different vehicle models and different parts based on the calculated expected value of the extended warranty cost.

[0010] In a preferred embodiment, S6 is specifically as follows: Establish a mapping relationship between the extended warranty service cost and the insurance actuarial rules based on the extended warranty cost decision matrix for different vehicle models and different parts; Calculate the initial value of the extended warranty service cost pricing for each vehicle model and each part according to the preset claim ratio limit, risk premium level, and profit margin ratio in the insurance actuarial rules; Use the insurance actuarial rules to perform actuarial verification on the initial value of the extended warranty service cost pricing to determine the final value of the extended warranty service cost that meets the insurance actuarial rules; Generate an extended warranty service pricing strategy for different vehicle models and different parts according to the final value of the extended warranty service cost.

[0011] On the other hand, the present invention provides an automotive extended warranty service cost optimization system based on big data analysis, including: Fault extraction module: perform semantic parsing on the vehicle historical maintenance record data to generate a fault feature vector; perform time series alignment processing on the real-time in-vehicle diagnostic data stream and output the time series fault feature; Mode division module: based on the fault feature vector and the time series fault feature, divide the high-frequency fault mode and the low-frequency fault mode, and generate a high-frequency fault probability distribution model and a low-frequency fault risk factor set; Cost prediction module: according to the high-frequency fault probability distribution model, calculate the expected maintenance cost in the high-frequency fault scenario through Monte Carlo simulation; Topological analysis module: perform topological correlation analysis on the low-frequency fault risk factor set for sparse fault events, identify the component-level fault chain reaction path, and generate a fault propagation network for low-frequency and high-cost events; Matrix generation module: based on the expected maintenance cost and the fault propagation network, optimize the output of the extended warranty cost decision matrix through a risk loss function; Pricing calculation module: according to the extended warranty cost decision matrix, call the preset insurance actuarial rule library to generate an extended warranty service pricing strategy.

[0012] The technical effects and advantages of an automotive extended warranty service cost optimization method and system based on big data analysis of the present invention: By performing semantic parsing on the historical vehicle maintenance record data and extracting fault feature vectors, and combining with real-time on-vehicle diagnostic data streams to construct time-series fault features, high-precision structured modeling of fault information is achieved; by dividing high-frequency and low-frequency fault modes, respectively constructing a high-frequency fault probability distribution model and a low-frequency fault risk factor set, typical faults and sparse high-loss events can be comprehensively covered; the Monte Carlo simulation method is used to predict the cost expectation of high-frequency faults, and the propagation path of low-frequency high-cost faults is identified through topological correlation analysis, effectively improving the modeling ability for complex risk structures; a risk loss function is constructed and the output extended warranty cost decision matrix is optimized, realizing fine control of extended warranty pricing factors; the insurance actuarial rule library is called to generate a differentiated extended warranty service pricing strategy, ensuring the applicability of the extended warranty service, improving the accuracy and robustness of pricing, and realizing the intelligence and dynamic optimization of the extended warranty service. Brief Description of the Drawings

[0013] Figure 1 Schematic diagram of an automobile extended warranty service cost optimization method based on big data analysis according to the present invention; Figure 2 Schematic diagram of the structure of an automobile extended warranty service cost optimization system based on big data analysis according to the present invention. Detailed Description of the Embodiments

[0014] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0015] Embodiment 1

[0016] Figure 1 An automobile extended warranty service cost optimization method based on big data analysis according to the present invention is given, which includes the following steps: S1: Perform semantic parsing on the historical vehicle maintenance record data to generate fault feature vectors; perform time-series alignment processing on the real-time on-vehicle diagnostic data streams and output time-series fault features; S2: Based on the fault feature vectors and time-series fault features, divide high-frequency fault modes and low-frequency fault modes, and generate a high-frequency fault probability distribution model and a low-frequency fault risk factor set; S3: According to the high-frequency fault probability distribution model, calculate the expected maintenance cost in the high-frequency fault scenario through Monte Carlo simulation; S4: For the low-frequency fault risk factor set, perform topological correlation analysis on sparse fault events, identify the component-level fault chain reaction path, and generate the fault propagation network of low-frequency high-cost events. S5: Optimize the output extended warranty cost decision matrix through a risk loss function based on the expected maintenance cost and the fault propagation network; S6: According to the extended warranty cost decision matrix, call the preset actuarial rule library to generate an extended warranty service pricing strategy.

[0017] S1: Perform semantic parsing on the vehicle historical maintenance record data to generate a fault feature vector; perform time series alignment processing on the real-time on-vehicle diagnostic data stream and output the time series fault features, including: Collect the vehicle historical maintenance record data; The vehicle historical maintenance record data includes all maintenance activity information that occurred during the extended warranty service period of the vehicle. The maintenance activity information includes, but is not limited to, the vehicle part maintenance time, maintenance location, fault phenomenon description, maintenance treatment method, maintenance cost, types of replaced parts during maintenance, and maintenance reasons. The acquisition channels of the vehicle historical maintenance record data include official maintenance record sheets given by the vehicle service center, vehicle maintenance invoices, spreadsheets exported from the vehicle maintenance registration system, and maintenance reports. The acquisition methods of the vehicle historical maintenance record data include connecting to the maintenance database to download the maintenance record sheet or importing the maintenance record through manual entry and electronic scanning.

[0018] For example, during the extended warranty period of a certain vehicle, there was an abnormal engine vibration problem. The vehicle owner sent the vehicle to the maintenance center for repair. The maintenance center recorded information such as the engine fault phenomenon, diagnostic results, replaced engine shock mounts, and maintenance costs, forming a complete maintenance record sheet. The maintenance record sheet is the vehicle historical maintenance record data.

[0019] Use natural language processing algorithms to perform semantic parsing and feature extraction on the vehicle historical maintenance record data to generate a fault feature vector; Adopt natural language processing algorithms to perform semantic parsing on the vehicle part maintenance time, maintenance location, fault phenomenon description, maintenance treatment method, maintenance cost, types of replaced parts during maintenance, and maintenance reasons described in text form in the vehicle historical maintenance record data, and convert the text description into a structured data form. The processing methods of semantic parsing include: performing word segmentation on the text in the vehicle historical maintenance record data, and constructing a semantic network using the co-occurrence relationship between words; identifying and classifying the fault phenomenon, maintenance measures, and maintenance parts based on the semantic network; based on the classified information, constructing a fault feature vector through feature extraction methods, including fault type features, maintenance measure features, part features, and maintenance cost features, etc.

[0020] Collect the real-time on-vehicle diagnostic data stream; Real-time on-vehicle diagnostic data stream refers to various vehicle operation parameter data collected in real time during the normal operation of a vehicle, including vehicle engine speed parameters, vehicle speed parameters, engine load parameters, coolant temperature parameters, intake manifold pressure parameters, vehicle fuel injection quantity parameters, and vehicle on-board diagnostic system fault codes. The collection of real-time on-vehicle diagnostic data stream is achieved by connecting a data acquisition device to the on-vehicle diagnostic interface of the vehicle. The data acquisition device continuously collects vehicle operation parameters at a set time frequency and outputs them in the form of a continuous data stream in real time. In practical applications, the data acquisition device includes a dedicated on-vehicle diagnostic data acquisition instrument and a real-time data transmission terminal.

[0021] Perform time series alignment processing based on a time window on the real-time on-vehicle diagnostic data stream to generate vehicle time series fault features; Normalize the real-time on-vehicle diagnostic data stream according to a unified time frequency, divide and segment the original real-time data stream based on a specific time interval window, and align the continuous vehicle operation parameter data in time series. The time series alignment processing method is as follows: Define a time window with a fixed duration, average and normalize the data within the window for the real-time on-vehicle diagnostic data stream, and process the irregular data stream into standardized data segments with a fixed time interval; Extract the corresponding vehicle operation state features in each standardized data segment, including state change trend features, stability features, and abnormal state occurrence features, to form vehicle time series fault features.

[0022] S2: Based on the fault feature vector and the time series fault features, divide the high-frequency fault modes and low-frequency fault modes, and generate a high-frequency fault probability distribution model and a set of low-frequency fault risk factors, including: Analyze the fault feature vector and the vehicle time series fault features through a deep neural network algorithm to generate the classification result of the fault modes of automotive components; Input the fault feature vector and the vehicle time series fault features into the deep neural network algorithm. The deep neural network algorithm includes an input layer, a hidden layer, and an output layer. Among them, the hidden layer is set with multiple network structure units for extracting complex non-linear features. The fault feature vector and the vehicle time series fault features are first input into the input layer, and the deep neural network algorithm performs layer-by-layer non-linear mapping on the input fault feature vector and vehicle time series fault features. Through the forward propagation calculation of the neural network algorithm, the node values of the input layer and the hidden layer are calculated in turn, and are processed by the activation function. Finally, the output layer outputs the classification result of the fault modes of automotive components. The activation function of the neural network algorithm adopts a non-linear function: Matrix operations and activation function transformations are performed on the network parameters of each layer of the input fault feature vector and vehicle time series fault features to determine the category of the fault modes of automotive components.

[0023] Set the fault frequency threshold according to the classification results of automotive component fault modes, and classify the automotive component fault modes into high-frequency fault modes and low-frequency fault modes based on the fault frequency threshold; According to the proportional relationship between the number of occurrences of various automotive component fault modes recorded in the classification results of automotive component fault modes and the total number of fault records, calculate the actual occurrence frequency corresponding to each fault mode. Set the threshold for dividing high-frequency fault modes and low-frequency fault modes: Sort all automotive component fault modes in descending order of frequency, and determine a frequency critical value that can distinguish the overall fault mode set into high-frequency and low-frequency categories. The automotive component fault modes with frequencies higher than the frequency critical value are classified as high-frequency fault modes, and the automotive component fault modes with frequencies lower than or equal to the frequency critical value are classified as low-frequency fault modes.

[0024] Based on the high-frequency fault modes, construct a high-frequency fault probability distribution model through probability statistical methods; Use the automotive component fault data corresponding to the high-frequency fault modes as input, and use probability statistical methods to statistically analyze the occurrence frequencies of various high-frequency fault modes, and calculate the occurrence probability of each high-frequency fault mode. The calculation process of the probability statistical method is as follows: Count the number of occurrences of each high-frequency fault mode in the entire high-frequency fault mode set, and then divide by the total number of fault records in the high-frequency fault mode set to obtain the fault occurrence probability corresponding to each high-frequency fault mode. The set of all high-frequency fault mode probabilities constitutes the high-frequency fault probability distribution model.

[0025] Based on the low-frequency fault modes, extract low-frequency fault risk characteristics through statistical analysis methods to generate a low-frequency fault risk factor set; Use statistical analysis methods to statistically analyze the costs, repair times, influence ranges, and occurrence of chain reactions of each automotive component fault in the low-frequency fault modes, and determine the risk characteristics in each low-frequency fault mode; The statistical analysis method is specifically as follows: Summarize the repair cost data of each low-frequency fault mode and the data on changes in vehicle operating states during fault occurrences, and use the risk value calculation method to calculate the potential risk characteristics of each low-frequency fault mode. The risk value calculation method includes the calculation of the average repair cost per single fault, the calculation of the proportion of the number of occurrences of fault chain reactions, and the calculation of the probability that the fault repair time exceeds a specific duration; Finally, comprehensively organize the risk characteristics of all low-frequency fault modes into a low-frequency fault risk factor set.

[0026] S3: According to the high-frequency fault probability distribution model, calculate the expected repair cost in the high-frequency fault scenario through Monte Carlo simulation, including: Establish a stochastic simulation model for high-frequency fault scenarios of automotive components based on the Monte Carlo simulation method; The Monte Carlo simulation method is a computational method based on the principle of random sampling. The high-frequency failure probabilities of each automotive component in the high-frequency failure probability distribution model are used as input data. A random simulation model is established: for the high-frequency failure of each automotive component, a random triggering mechanism is set according to the corresponding occurrence probability of the high-frequency failure. The random triggering mechanism determines whether the corresponding failure occurs in each simulation through a random number generation method during the simulation process. The random number generation method is as follows: within a preset probability space, a random value is uniformly and randomly generated. If the random value is within the range corresponding to the failure probability, the high-frequency failure scenario of the corresponding automotive component is triggered. After the random simulation model is set up, it can randomly simulate the high-frequency failure scenarios of automotive components.

[0027] For example, the failure of a certain automotive air-conditioning compressor corresponds to a relatively high occurrence probability in the high-frequency failure probability distribution model. A random triggering mechanism for the air-conditioning compressor failure is set in the random simulation model. During each simulation process, a random number is uniformly and randomly generated. If the random number falls within the range of the air-conditioning compressor failure probability, it is determined that the compressor failure is triggered during this simulation process.

[0028] The random simulation model for the high-frequency failure scenarios of automotive components is iteratively run multiple times according to a preset number of simulation times. Set the number of iterative runs of the simulation model. Start the random simulation model for simulation according to the set number of iterative runs one by one. Each iterative simulation is independent; each simulation is based on the random triggering mechanism in the random simulation model to generate high-frequency failure events of automotive components; after multiple iterative simulation runs, a set of simulation result data is formed.

[0029] For example, the random simulation model sets a certain number of simulation times. Each simulation simulates the high-frequency failure scenarios that the vehicle may encounter; for example, a certain scale of the number of simulation times is set, and each simulation is independently run. Each run generates automotive component failure events according to the random triggering mechanism, and the simulation data is accumulated.

[0030] During each iterative run, a failure event that conforms to the high-frequency failure probability distribution model is randomly generated, and the single repair cost of each failure event is determined. During each simulation run, the high-frequency failure events of automotive components that occur are determined according to the random triggering mechanism in the random simulation model; according to the historical repair cost data corresponding to the high-frequency failure of the automotive components in the vehicle historical repair record data, the single repair cost of each failure event is statistically calculated; the calculation method of the single repair cost is as follows: the repair cost data of the same type of failure in the vehicle historical repair record data is statistically analyzed, and the average value of the historical repair cost data is calculated to determine the repair cost of each simulation failure event.

[0031] Statistically analyze the single repair cost for all iterations and calculate the expected repair cost for the high-frequency failure scenarios of automotive parts; Collate and summarize all single repair cost data sets; then calculate the expected repair cost for the high-frequency failure scenarios of automotive parts. The formula for calculating the expected repair cost is as follows: Add up the values of all single repair cost data to obtain the total repair cost, and then divide the total repair cost by the total number of iterative simulation runs to get the average repair cost for the high-frequency failure scenarios of automotive parts, which is the expected repair cost.

[0032] S4: For the set of low-frequency failure risk factors, conduct a topological correlation analysis on sparse failure events, identify the component-level failure chain reaction paths, and generate a failure propagation network for low-frequency high-cost events, including: Based on the set of low-frequency failure risk factors, construct a topological correlation model for low-frequency failures of automotive parts using the spatial enhanced topology algorithm; The set of low-frequency failure risk factors includes the single failure average repair cost characteristics of automotive parts, the proportion characteristics of the number of failure chain reactions, the probability characteristics of the repair time exceeding a specific duration, and the vehicle operating state change characteristics during failure. The implementation method of the spatial enhanced topology algorithm is as follows: Take each low-frequency failure of automotive parts as a node, and use the association relationships existing between the relevant low-frequency failures of automotive parts in the set of low-frequency failure risk factors as the edges connecting the nodes; Add a spatial enhancement mechanism to the network structure and adjust the weights of the network edges based on the risk characteristic values of the nodes. The calculation method for the weights of the network edges is as follows: Based on the risk characteristic values of the single failure average repair cost characteristics, the proportion characteristics of the number of failure chain reactions, and the repair time characteristics corresponding to the nodes, perform a comprehensive operation. The weight value is equal to the sum of the products of the above three risk characteristic values multiplied by their corresponding weight coefficients respectively; By adjusting the weights of the network edges, a topological correlation model is formed. The topological correlation model can reflect the spatial correlation characteristics between low-frequency failures of automotive parts.

[0033] Analyze the mutual association relationships between low-frequency failures of automotive parts using the topological correlation model of low-frequency failures of automotive parts to determine the association strength between low-frequency failures of automotive parts; Conduct an association strength analysis using the edge weight values between nodes in the topological correlation model; The calculation method for the association strength is as follows: For any two connected low-frequency failure nodes of automotive parts in the topological correlation model, based on the weight value of the connecting edge, determine the association strength between the low-frequency failure nodes; When calculating the association strength, perform a proportional normalization process based on the weight value of the edge to calculate the association strength between each pair of nodes. Specifically: Divide the weight value of a single connecting edge by the sum of the weight values of all connecting edges to obtain the association strength; After calculating the association strength between all nodes, form an association strength matrix for low-frequency failures.

[0034] Identify the chain reaction paths of low-frequency failures of automotive parts according to the correlation strength, and generate a fault propagation network for low-frequency, high-cost fault events of automotive parts; Determine the chain reaction relationship between nodes according to the correlation strength matrix; The chain reaction path identification method is as follows: Connect the paths of nodes with correlation strength values higher than a specific threshold in the correlation strength matrix; The path connection method is to determine two low-frequency failure nodes of automotive parts with correlation strength higher than the specific threshold as the starting node and the ending node in the chain reaction path, and the path direction extends from the node with a higher correlation strength value to the node with a lower correlation strength value; Sequentially determine the chain reaction paths between multiple nodes, and the path length is determined in sequence according to the magnitude of the correlation strength value; Through the above method, identify the chain reaction paths of low-frequency failures of automotive parts, and form the fault propagation network of low-frequency, high-cost fault events of automotive parts as a whole.

[0035] S5: Based on the expected maintenance cost and the fault propagation network, optimize and output the extended warranty cost decision matrix through the risk loss function, including: Construct a risk loss function based on the expected maintenance cost of high-frequency fault scenarios of automotive parts and the fault propagation network of low-frequency, high-cost fault events of automotive parts; The construction method of the risk loss function is as follows: Determine the input parameters of the risk loss function, including the expected maintenance cost value of high-frequency fault scenarios of automotive parts and the chain reaction paths and the maintenance cost data on the chain reaction paths in the propagation network of low-frequency, high-cost fault events of automotive parts; The expression of the risk loss function adopts the method of adding the products of the maintenance costs of each scenario and the corresponding occurrence probabilities. Specifically: Multiply the expected maintenance cost value of the high-frequency fault scenario of automotive parts by the corresponding occurrence probability value to obtain the risk loss term of the high-frequency fault scenario; Multiply the single maintenance cost value on each low-frequency fault path in the fault propagation network by the corresponding occurrence probability value of each path to obtain the risk loss term of each low-frequency fault path; Then add the risk loss term of the high-frequency fault scenario and the risk loss terms of each low-frequency fault path to obtain the expression of the risk loss function.

[0036] Use the risk loss function to optimize and analyze the risk weight coefficients of high-frequency fault scenarios of automotive parts and low-frequency, high-cost fault events of automotive parts; The method for optimizing the risk weight coefficient is as follows: Adjust the risk weight coefficient in the risk loss function and calculate the risk loss values under different combinations of weight coefficients, and finally determine the combination of weight coefficients that minimizes the risk loss value. Specifically: Initially set different combinations of risk weight coefficients, where the combination of weight coefficients represents the proportions of high-frequency failure scenarios and low-frequency failure events in the total risk loss function respectively; substitute them into the risk loss function for calculation to obtain multiple risk loss results; then, by comparing the calculation results, determine a set of risk weight coefficients that can minimize the value of the risk loss function.

[0037] According to the optimized risk weight coefficient, calculate the expected values of the extended warranty costs corresponding to the high-frequency failure scenarios of automotive parts and the low-frequency high-cost failure events of automotive parts; The calculation method of the expected value of the extended warranty cost is as follows: Multiply the optimized risk weight coefficient of the high-frequency failure scenario by the expected maintenance cost value of the high-frequency failure scenario of automotive parts; The calculation method of the expected value of the extended warranty cost for the low-frequency high-cost failure event of automotive parts is: Multiply the optimized risk weight coefficient of the low-frequency failure path by the product of the maintenance cost and probability of each low-frequency failure path in the failure propagation network, and add up the cost expected values of all low-frequency failure paths. Add the above two cost expected values to calculate the overall expected value of the extended warranty cost.

[0038] Based on the calculated expected values of the extended warranty costs, generate an extended warranty cost decision matrix for different vehicle models and different parts; The method for generating the extended warranty cost decision matrix is as follows: Classify and organize the expected values of the extended warranty costs in two dimensions according to the vehicle models and types of parts, and establish the structure of the cost decision matrix; Fill in the expected values of the extended warranty costs for different vehicle models and different parts into the corresponding matrix cells in the cost decision matrix; The horizontal axis of the matrix is the types of automotive parts, and the vertical axis is the vehicle models; The value in the cell is the calculated expected value of the extended warranty cost.

[0039] S6: According to the extended warranty cost decision matrix, call the preset insurance actuarial rule library to generate an extended warranty service pricing strategy, including: Based on the extended warranty cost decision matrix for different vehicle models and different parts, establish a mapping relationship between the extended warranty service cost and the insurance actuarial rules; The horizontal axis of the extended warranty cost decision matrix represents different types of automotive parts, including but not limited to engines, transmissions, air conditioning compressors, and vehicle electronic control units. The vertical axis represents different automotive models, such as sedans, sport utility vehicles, and trucks. The actuarial rules include the ratio limits for vehicle part repair claims, the method for determining risk premiums, and the setting criteria for profit margins. The method for establishing the mapping relationship between the extended warranty service cost and the actuarial rules is as follows: Each cost value in the extended warranty cost decision matrix is corresponded one by one with the corresponding rules of the actuarial rules, including associating the extended warranty cost values of each model and each part with the claim rate limits, risk premium levels, and profit margin ratios set in the actuarial rules one by one, forming a one-to-one cost-rule mapping relationship table. The mapping relationship shows that the extended warranty cost of each part of each model corresponds to the claim rate limit range, risk premium standard, and profit margin requirement respectively.

[0040] Calculate the initial value of the extended warranty service cost pricing for each model and each part according to the claim rate limits, risk premium levels, and profit margin ratios preset in the actuarial rules; The calculation method of the initial value of the extended warranty service cost pricing is as follows: Determine the basic cost value of the extended warranty service according to the claim rate limit in the actuarial rules; The application method of the claim rate limit is to multiply the repair cost of each part by the ratio set by the claim rate limit to calculate the basic cost value. Adjust the basic cost value using the risk premium level. The specific calculation method for adjusting the risk premium level is to multiply the basic cost value by the ratio determined by the risk premium level to calculate the risk premium cost value. Based on the risk premium cost value, calculate the profit cost value according to the profit margin ratio: Multiply the risk premium cost value by the profit margin ratio again; Finally, add the basic cost value, risk premium cost value, and profit cost value to obtain the initial value of the extended warranty service cost pricing for each model and each part.

[0041] Use the actuarial rules to perform actuarial verification on the initial value of the extended warranty service cost pricing to determine the final value of the extended warranty service cost that meets the actuarial rules; The method of actuarial verification is as follows: Substitute the calculated initial value of the extended warranty service cost pricing for each model and each part into the actuarial rules one by one for item-by-item verification. The verification content includes whether the initial value of the extended warranty cost pricing meets the claim rate limit range, the control standard of the risk premium, and the limit standard of the profit margin ratio. Verification process: Compare the initial value of the extended warranty cost pricing with the claim rate limit standard, risk premium level, and profit margin ratio item by item. If there is an initial pricing value that does not meet the actuarial rule standard, adjust it: Recalculate according to the standard ratio limited by the actuarial rules until all actuarial rule limits are met. After the actuarial verification is completed, form the final value of the extended warranty service cost within the scope of the actuarial rules.

[0042] Generate an extended warranty service pricing strategy for different vehicle models and different components according to the final determined value of the extended warranty service cost; The method for generating the extended warranty service pricing strategy is as follows: taking the final determined value of the extended warranty service cost as the core, determine the corresponding market price strategy according to the market positioning of different vehicle models and the risk characteristic differences of different components. The market price strategy includes market price determination, differential pricing, and promotional pricing schemes. The method for market price determination is: combine the final determined value of the extended warranty service cost with the market demand elasticity coefficient to determine the price value suitable for market sales; the method for differential pricing is: formulate differential pricing values for specific consumer groups of different vehicle models and different components; the way to formulate the promotional pricing scheme is to make price discounts or special offers for the extended warranty service pricing of some components according to the market promotion plan during a specific period, and finally form an extended warranty service pricing strategy for different vehicle models and different components.

[0043] Embodiment 2

[0044] The difference between Embodiment 2 and Embodiment 1 of the present invention is that this embodiment introduces an optimization system for the extended warranty service cost of automobiles based on big data analysis.

[0045] Figure 2 The structural schematic diagram of an optimization system for the extended warranty service cost of automobiles based on big data analysis according to the present invention is given. An optimization system for the extended warranty service cost of automobiles based on big data analysis includes: Fault extraction module: perform semantic parsing on the historical vehicle repair record data to generate fault feature vectors; perform time series alignment processing on the real-time in-vehicle diagnostic data stream and output time series fault features; Mode division module: based on the fault feature vectors and time series fault features, divide high-frequency fault modes and low-frequency fault modes, and generate a high-frequency fault probability distribution model and a low-frequency fault risk factor set; Cost prediction module: according to the high-frequency fault probability distribution model, calculate the expected repair cost in the high-frequency fault scenario through Monte Carlo simulation; Topological analysis module: for the low-frequency fault risk factor set, perform topological correlation analysis on sparse fault events, identify component-level fault chain reaction paths, and generate a fault propagation network for low-frequency and high-cost events; Matrix generation module: based on the expected repair cost and the fault propagation network, optimize and output an extended warranty cost decision matrix through a risk loss function; Pricing calculation module: according to the extended warranty cost decision matrix, call a preset insurance actuarial rule library to generate an extended warranty service pricing strategy.

[0046] The above formulas are all dimensionless and take their numerical values for calculation. The formulas are obtained by collecting a large amount of data for software simulation to get a formula that is closest to the actual situation. The preset parameters and threshold selection in the formulas are set by those skilled in the art according to the actual situation.

[0047] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by wired (such as infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that can be accessed by a computer, or a data storage device such as a server or data center that contains one or more sets of available media. The available media can be magnetic media (such as floppy disks, hard disks, magnetic tapes), optical media (such as DVDs), or semiconductor media. The semiconductor media can be a solid-state drive.

[0048] Those of ordinary skill in the art can realize that the modules and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0049] Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working processes of the systems, devices, and modules described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0050] In several embodiments provided in the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the modules is only a logical function division. In actual implementation, there may be other division methods. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces. The indirect couplings or communication connections of devices or modules can be in electrical, mechanical, or other forms.

[0051] The modules described as separate components may or may not be physically separated. The components shown as modules may or may not be physical modules. They can be located in one place or distributed to multiple network modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0052] In addition, in each embodiment of the present application, the functional modules can be integrated into one processing module, or each module can exist physically alone, or two or more modules can be integrated into one module.

[0053] If the above functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or this 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 for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0054] The above is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present application can easily think of changes or substitutions, which should all be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

[0055] Finally, the above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. An optimization method for the cost of vehicle extended warranty services based on big data analysis, characterized in that, It includes the following steps: S1: Perform semantic parsing on the historical vehicle repair record data to generate a fault feature vector; perform time series alignment processing on the real-time on-vehicle diagnostic data stream and output the time series fault features; S2: Based on the fault feature vector and the time series fault features, divide the high-frequency fault modes and low-frequency fault modes, and generate a high-frequency fault probability distribution model and a set of low-frequency fault risk factors; S3: According to the high-frequency fault probability distribution model, calculate the expected repair cost in the high-frequency fault scenario through Monte Carlo simulation; S4: For the set of low-frequency fault risk factors, perform topological correlation analysis on the sparse fault events, identify the component-level fault chain reaction paths, and generate a fault propagation network for low-frequency and high-cost events; S5: Based on the expected repair cost and the fault propagation network, optimize and output the extended warranty cost decision matrix through a risk loss function; S6: According to the extended warranty cost decision matrix, call the preset actuarial rule library to generate an extended warranty service pricing strategy.

2. The method for optimizing the cost of vehicle extended warranty service based on big data analysis according to claim 1, wherein, S1 specifically is: Collect the historical vehicle repair record data; Use natural language processing algorithms to perform semantic parsing and feature extraction on the historical vehicle repair record data to generate a fault feature vector; Collect the real-time on-vehicle diagnostic data stream; Perform time series alignment processing based on a time window on the real-time on-vehicle diagnostic data stream to generate the vehicle time series fault features.

3. The method for optimizing the cost of vehicle extended warranty services based on big data analysis according to claim 2, wherein S2 specifically is: Analyze the fault feature vector and the vehicle time series fault features through a deep neural network algorithm to generate the classification results of the automotive component fault modes; Set a fault frequency threshold according to the classification results of the automotive component fault modes, and divide the automotive component fault modes into high-frequency fault modes and low-frequency fault modes based on the fault frequency threshold; Based on the high-frequency fault modes, construct a high-frequency fault probability distribution model through probability statistical methods; Based on the low-frequency fault modes, extract low-frequency fault risk features through statistical analysis methods to generate a set of low-frequency fault risk factors.

4. The method for optimizing the cost of an extended warranty service for an automobile based on big data analysis according to claim 3, wherein, S3 specifically is: Establish a stochastic simulation model for high-frequency fault scenarios of automotive components based on the Monte Carlo simulation method; Perform multiple iterative runs on the stochastic simulation model for high-frequency fault scenarios of automotive components according to the preset number of simulation times; During each iterative run, randomly generate fault occurrence events that conform to the high-frequency fault probability distribution model, and determine the single repair cost of each fault event; Statistically analyze the single repair costs of all iterative runs to calculate the expected repair cost in the high-frequency fault scenario of automotive components.

5. A method for optimizing the cost of vehicle extended warranty services based on big data analysis according to claim 4, characterized in that, S4 specifically is: Based on the set of low-frequency fault risk factors, construct a topological correlation model for low-frequency faults of automotive components using a spatial enhanced topology algorithm; Use the topological correlation model for low-frequency faults of automotive components to analyze the mutual correlation relationships between low-frequency faults of automotive components, and determine the correlation strength between low-frequency faults of automotive components; Identify the chain reaction paths for low-frequency faults of automotive components according to the correlation strength, and generate a fault propagation network for low-frequency and high-cost fault events of automotive components.

6. The method for optimizing the cost of vehicle extended warranty services based on big data analysis according to claim 5, wherein S5 specifically is: Based on the expected repair cost in the high-frequency fault scenario of automotive components and the fault propagation network for low-frequency and high-cost fault events of automotive components, construct a risk loss function; Optimize and analyze the risk weight coefficients of high-frequency failure scenarios of automotive parts and low-frequency and high-cost failure events of automotive parts using a risk loss function; Calculate the expected extended warranty cost corresponding to the high-frequency failure scenario of automotive parts and the low-frequency and high-cost failure event of automotive parts according to the optimized risk weight coefficient; Generate an extended warranty cost decision matrix for different vehicle models and different parts based on the calculated expected extended warranty cost; 7. The method for optimizing the cost of vehicle extended warranty services based on big data analysis according to claim 6, characterized in that S6, specifically: Based on the extended warranty cost decision matrix for different vehicle models and different parts, establish a mapping relationship between the extended warranty service cost and the insurance actuarial rules; Calculate the initial value of the extended warranty service cost pricing for each vehicle model and each part according to the preset loss ratio limit, risk premium level, and profit margin ratio in the insurance actuarial rules; Use the insurance actuarial rules to perform actuarial verification on the initial value of the extended warranty service cost pricing to determine the final value of the extended warranty service cost that meets the insurance actuarial rules; Generate an extended warranty service pricing strategy for different vehicle models and different parts according to the final value of the extended warranty service cost; 8. An automobile extended warranty service cost optimization system based on big data analysis, which is used to implement the method for optimizing the cost of an automobile extended warranty service according to any one of claims 1-7, characterized in that, Including: Fault extraction module: Perform semantic parsing on the vehicle historical maintenance record data to generate a fault feature vector; Perform time series alignment processing on the real-time on-vehicle diagnostic data stream and output the time series fault features; Mode division module: Based on the fault feature vector and the time series fault features, divide the high-frequency and low-frequency fault modes, and generate a high-frequency fault probability distribution model and a low-frequency fault risk factor set; Cost prediction module: According to the high-frequency fault probability distribution model, calculate the expected maintenance cost in the high-frequency failure scenario through Monte Carlo simulation; Topology analysis module: For the low-frequency fault risk factor set, perform topological correlation analysis on the sparse fault events, identify the component-level fault chain reaction path, and generate a fault propagation network for low-frequency and high-cost events; Matrix generation module: Based on the expected maintenance cost and the fault propagation network, optimize and output the extended warranty cost decision matrix through a risk loss function; Pricing calculation module: According to the extended warranty cost decision matrix, call the preset insurance actuarial rule library to generate an extended warranty service pricing strategy.

Images