Identification, early warning and legal reminder methods based on big data

Abstract

This invention provides a big data-based identification, early warning, and legal warning method. The method involves real-time collection of multidimensional user behavioral data, including identity characteristics, historical transaction prices, and real-time environmental variables; performing spatiotemporal correlation processing on the data to generate a spatiotemporally aligned feature tensor; performing differentiated weighted fusion of the feature tensors based on user price sensitivity levels to generate a classification feature matrix; calculating a dynamic deviation index and generating a comprehensive price rationality score, with low scores triggering early warnings; inputting this into a legal clause matching engine to generate early warning reports; and implementing multi-level intervention strategies based on the warning level. This invention can help users identify potential price discrepancies and provide technical support for related platforms, enabling more accurate detection and resolution of price anomalies.

Description

Technical Field

[0001] The present invention relates to the field of big data analysis and artificial intelligence technology, and more specifically, to a method for identifying, warning, and providing legal reminders based on big data. Background Art

[0002] With the booming development of e-commerce, online shopping has become an integral part of daily life. Merchants are leveraging big data analytics to optimize pricing strategies, enhance user experience, and improve operational efficiency. However, this data-driven pricing strategy has also raised concerns. For example, merchants are exploiting consumer personal information to impose higher prices or other unfavorable transaction conditions on repeat customers or high-spending users in order to maximize profits. This behavior not only harms the legitimate rights and interests of consumers.

[0003] Existing pricing strategies and regulatory mechanisms rely primarily on merchant self-regulation and subsequent consumer complaints. Merchants often adjust prices based on user purchase history, browsing behavior, and other data, but these adjustments often lack transparency and are difficult for consumers to detect. Handling this feedback often struggles to identify and resolve issues promptly due to a lack of real-time monitoring and effective technical means. Furthermore, existing technical approaches primarily focus on user behavior analysis and price monitoring, but these methods can only detect price fluctuations and cannot accurately determine whether pricing is being misled, much less provide robust evidence or early warnings. Summary of the Invention

[0004] The present invention provides a method for identifying, warning and providing legal reminders based on big data, including:

[0005] S1. Real-time collection of multi-dimensional user behavior data, including user identity feature data, historical transaction price series, and real-time environmental variable data. The user identity feature data includes at least device fingerprint features, membership level tags, and spending capacity assessment values, which are calculated by weighting the user's historical order amount and consumption frequency.

[0006] S2. Performing spatiotemporal correlation processing on the multi-dimensional behavioral data to generate a spatiotemporally aligned feature tensor by detecting abnormal transaction time points and mapping user identities across platforms, wherein the cross-platform user identity mapping is based on dual verification of device fingerprint similarity and consumer behavior consistency;

[0007] S3. Based on the user price sensitivity classification results, perform differentiated weighted fusion on the feature tensor to generate a user group classification feature matrix, wherein the historical transaction price fluctuation threshold of the highly price-sensitive user group is reduced by 20%-30%;

[0008] S4. Calculate the dynamic deviation index of the real-time transaction price relative to the user's historical benchmark price, and integrate the real-time supply and demand index with the promotion intensity parameter to generate a comprehensive score for price rationality. When the score falls below the preset risk level, trigger an early warning mechanism;

[0009] S5. Input the warning scenario feature vector into a legal clause matching engine to generate a warning report containing the anti-price discrimination legal basis and penalty clauses. The legal clause matching engine is implemented by semantic analysis of legal provisions and similarity calculation of scenario feature vectors.

[0010] S6. Implement multi-level intervention strategies based on the warning report level, including pushing real-time price comparison information to user terminals, sending interface flow limiting instructions to the enterprise risk control system, and submitting risk filing data packages to the regulatory platform.

[0011] Furthermore, step S2 includes:

[0012] S21, using dynamic time window to detect abnormal transaction time points, when it is detected that the transaction price is within the time window, When the fluctuation range exceeds 3 times the standard deviation of the historical period, the original timestamp will be compensated and corrected. The correction formula is:

[0013] ;

[0014] in, represents the transaction price at time t, represents the average price during the window period, is the dynamic compensation coefficient and , is a symbolic function;

[0015] S22. Build a cross-platform user identity mapping graph, calculate the similarity of the device fingerprint hash value and the matching degree of the consumption behavior sequence, and generate a user identity confidence score. When the confidence score is greater than 0.8, it is determined to be the same user entity;

[0016] S23. Reconstruct the spatiotemporally aligned data into a three-dimensional feature tensor, the dimensions of which include the number of users, the number of time slices, and the number of feature channels, where the feature channels at least include price volatility, consumption frequency change, and environmental variable influencing factors.

[0017] Furthermore, the differentiated weighted fusion in step S3 includes:

[0018] S31. Calculate the user price sensitivity index by analyzing the user's price comparison response time in historical orders, coupon usage efficiency, and preference for alternative products. The preference for alternative products should be weighted at no less than 40%.

[0019] S32. Build a user group classification model, classifying the top 10% of users in terms of price sensitivity index as the highly sensitive group, the middle 30% as the moderately sensitive group, and the remainder as the low-sensitivity group. Different price fluctuation tolerance coefficients are set for different groups.

[0020] S33. Perform group weighted pooling processing on the feature tensor. The feature channel weights of the highly sensitive group are increased by 1.5-2 times, and the weights of the low-sensitive group are reduced to 0.7-0.9 times, thereby generating a classification feature matrix with group discrimination.

[0021] Furthermore, step S4 includes:

[0022] S41. Generate a user's historical benchmark price curve, using an exponential decay weighting method to integrate the user's past K transaction prices, with the transaction prices in the last 30 days accounting for no less than 60% of the weight;

[0023] S42. Calculate the dynamic price deviation index by comparing the absolute difference between the real-time transaction price and the benchmark price, and superimposing the influencing factors of the real-time supply and demand index deviation to generate a normalized deviation score. The specific formula is:

[0024] ;

[0025] in, is the base price, Indicates the deviation between the real-time supply and demand index and the historical benchmark value. is the benchmark value of the supply and demand index;

[0026] S43. Construct a price rationality scoring model. When the deviation index exceeds the tolerance threshold of the user's group, activate the price rationality degradation mechanism. Every 10-point drop in the score triggers a first-level warning.

[0027] Furthermore, the legal terms matching engine workflow in step S5 includes:

[0028] S51. Construct a legal knowledge graph, parse the relevant anti-price discrimination laws into structured (illegal behavior, identification requirements, penalty measures) triples, and map them into a high-dimensional feature space;

[0029] S52. Calculate the matching degree between the real-time scenario feature vector and the legal provision feature vector, focusing on the three core dimensions of price difference rationality, user identity relevance, and subjective intent determination, and set a dynamic weight coefficient for each dimension;

[0030] S53. Generate a legal basis report. When the matching degree exceeds 0.75, it will automatically link the precedents of similar cases in the past three years, and mark the median penalty amount and common defense reasons.

[0031] Furthermore, the method for calculating the user price sensitivity index also includes:

[0032] S311. Analyze the user's price comparison behavior characteristics, count the user's stay time and number of jumps to the price comparison page, and generate a price attention index;

[0033] S312. Calculate the coupon usage efficiency by comparing the difference in usage time after coupon redemption with the increase in order amount to assess the user's response speed to price incentives.

[0034] S313. Build an alternative product selection model to record the frequency of users switching between products in different price ranges. When the frequency is more than twice the average level of similar users, it is determined to be a high price sensitivity feature.

[0035] Furthermore, the calculation method of the supply and demand index deviation includes:

[0036] S421. Collect real-time market data, including median prices on competing platforms, inventory turnover rates, and search popularity indexes;

[0037] S422. Calculate the supply-demand balance coefficient:

[0038] ;

[0039] in, is the average price of competing products, is the real-time search volume, is the current inventory, is the supply-demand balance coefficient, is the mean search volume, For the price of products on this platform, is the average inventory level;

[0040] S423. Normalize the supply-demand balance coefficient to generate an impact factor in the range [0,1].

[0041] Furthermore, the method for constructing the legal text feature vector includes:

[0042] S511. Use natural language processing technology to parse legal texts and extract key elements of the elements of illegal behavior, including the subject of the behavior, subjective aspects, objective behavior, and damage results;

[0043] S512, mapping each element to a 64-dimensional semantic vector space, and calculating the association weights between elements through the attention mechanism;

[0044] S513. When generating the feature vector of the article, the penalty amount interval is discretized and encoded, and the limitation period is converted into an attenuation coefficient.

[0045] Furthermore, the multi-level intervention strategy in step S6 includes:

[0046] S61: At the high-risk warning level, execute simultaneously:

[0047] Real-time freezing of the price adjustment interface for suspected abnormal products

[0048] Show users the price comparison information flow of the product on different user terminals

[0049] Automatically generate electronic evidence packages containing timestamps, price comparison charts, and legal basis;

[0050] S62, medium-risk warning level, triggers:

[0051] Extend the price update review period to a minimum of 2 hours

[0052] Send a second confirmation request to the operator, and provide a justification for the price adjustment;

[0053] S63, low-risk warning level, execute:

[0054] Record abnormal transaction logs and mark suspected feature labels

[0055] Start tracking and monitoring of similar transaction behaviors for the next 7 days.

[0056] Furthermore, it also includes the model iterative optimization mechanism:

[0057] S71. Collect user feedback data and subsequent transaction behavior data after the warning is triggered, and construct a feedback training set;

[0058] S72. Optimize the parameters of the price rationality scoring model, focusing on adjusting the proportional relationship between the supply and demand index influencing factors and the user group weight coefficients. The optimization objective function is:

[0059] ;

[0060] in, 、 is the balance coefficient, represents the mean absolute error, To calculate the KL divergence of the warning distribution, is the model prediction value, is the true value, is the model prediction distribution, is the true distribution.

[0061] The above-described embodiments of the present invention have at least the following beneficial effects: the present invention can accurately identify and provide real-time early warnings for abnormal pricing behavior. By collecting multidimensional user behavior data in real time and performing spatiotemporal correlation processing, combined with differentiated weighted fusion based on user price sensitivity grading, it can accurately capture abnormal transaction characteristics and generate a highly discriminatory classification feature matrix. Based on this, a dynamic deviation index is calculated and a comprehensive price rationality score is generated. When the score falls below a preset risk level, an early warning mechanism is triggered, thereby promptly identifying potential abnormal pricing behavior. This provides users with a more transparent consumption environment and helps them make more informed consumer decisions. It also provides technical support for businesses and helps them optimize their pricing strategies. Furthermore, the present invention can also provide risk alerts for early warning scenarios. The early warning scenario feature vector is input into a legal clause matching engine. Through semantic analysis of legal text and similarity calculation of the scenario feature vector, an early warning report is generated containing the legal basis and reasonable explanations for the relevant legal clauses. This not only helps users protect their legitimate rights and interests when encountering abnormal pricing behavior, but also provides businesses with clear compliance guidance, improving their operational efficiency and promoting reasonable and fair pricing. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] The above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily apparent by reading the following detailed description with reference to the accompanying drawings, in which several embodiments of the present invention are shown by way of example and not limitation, in which:

[0063] Figure 1 A flowchart of a method for identifying, warning, and providing legal reminders based on big data is provided in accordance with an embodiment of the present invention. DETAILED DESCRIPTION

[0064] The principles and spirit of the present invention will be described below with reference to several exemplary embodiments. It should be understood that these embodiments are provided solely to enable those skilled in the art to better understand and implement the present invention, and are not intended to limit the scope of the present invention in any way. Rather, these embodiments are provided to make the present invention more thorough and complete, and to fully convey the scope of the present invention to those skilled in the art.

[0065] Those skilled in the art will appreciate that the embodiments of the present invention may be implemented as a system, apparatus, device, method, or computer program product. Therefore, the present invention may be implemented in the following forms: entirely in hardware, entirely in software (including firmware, resident software, microcode, etc.), or in a combination of hardware and software.

[0066] It should be noted that any number of elements in the drawings is for illustration only and not for limitation, and any naming is only for distinction and does not have any limiting meaning.

[0067] Reference below Figure 1 , Figure 1 This is a flow chart of a method for identifying, warning and providing legal reminders based on big data provided by one embodiment of the present invention. Figure 1 As shown, a method for identifying, warning and providing legal reminders based on big data includes:

[0068] S1. Real-time collection of multi-dimensional user behavior data, including user identity feature data, historical transaction price series, and real-time environmental variable data. The user identity feature data includes at least device fingerprint features, membership level tags, and spending capacity assessment values, which are calculated by weighting the user's historical order amount and consumption frequency.

[0069] S2. Performing spatiotemporal correlation processing on the multi-dimensional behavioral data to generate a spatiotemporally aligned feature tensor by detecting abnormal transaction time points and mapping user identities across platforms, wherein the cross-platform user identity mapping is based on dual verification of device fingerprint similarity and consumer behavior consistency;

[0070] S3. Based on the user price sensitivity classification results, perform differentiated weighted fusion on the feature tensor to generate a user group classification feature matrix, wherein the historical transaction price fluctuation threshold of the highly price-sensitive user group is reduced by 20%-30%;

[0071] S4. Calculate the dynamic deviation index of the real-time transaction price relative to the user's historical benchmark price, and integrate the real-time supply and demand index with the promotion intensity parameter to generate a comprehensive score for price rationality. When the score falls below the preset risk level, trigger an early warning mechanism;

[0072] S5. Input the warning scenario feature vector into a legal clause matching engine to generate a warning report containing the anti-price discrimination legal basis and penalty clauses. The legal clause matching engine is implemented by semantic analysis of legal provisions and similarity calculation of scenario feature vectors.

[0073] S6. Implement multi-level intervention strategies based on the warning report level, including pushing real-time price comparison information to user terminals, sending interface flow limiting instructions to the enterprise risk control system, and submitting risk filing data packages to the regulatory platform.

[0074] It should be noted that the present invention proposes a method for identification, early warning and legal reminder based on big data. The core of this method is to build a comprehensive user behavior portrait by collecting multi-dimensional user behavior data in real time, including user identity feature data, historical transaction price series and real-time environmental variable data. User identity feature data at least includes device fingerprint features, member level labels and consumption capacity evaluation values, where the consumption capacity evaluation value is obtained by weighted calculation of the user's historical order amount and consumption frequency. This indicator can effectively reflect the user's consumption level and purchasing power. The collection of multi-dimensional behavior data provides a rich data basis for subsequent analysis and early warning, enabling the system to evaluate and judge the user's transaction behavior from multiple angles.

[0075] Specifically, the device fingerprint feature within user identity data refers to a unique identifier generated by collecting multi-dimensional information such as the user's device hardware, software configuration, and network environment. This unique identifier is used to identify the behavior of the same device across different platforms. Membership tiers are categorized based on metrics such as spending amount and frequency, such as regular, gold, and platinum. Different membership tiers typically correspond to different benefits and services. The spending capacity assessment is calculated by weighting the user's historical order amount and spending frequency. The weighting can be adjusted based on actual circumstances. For example, for users who frequently consume but spend relatively low amounts each time, the weighting of spending frequency can be appropriately increased. The historical transaction price series refers to the amount of all transactions a user has made over a certain period of time. Analyzing this data can help understand a user's price sensitivity and purchasing habits. Real-time environmental variable data includes current market supply and demand, promotional activities, and other factors, all of which influence a user's transaction price.

[0076] Preferably, when collecting multi-dimensional behavioral data of users, the source and type of data can be further refined. For example, in the process of collecting device fingerprint features, the user's geographic location information can be combined to analyze the IP address, GPS location and other data of the user's device to determine whether the user is trading in a specific area, thereby eliminating some abnormal trading behaviors. For historical transaction price series, a time window can be set, such as transaction records in the past year or six months, in order to more accurately reflect the user's recent consumption behavior. In terms of collecting real-time environmental variable data, in addition to market supply and demand conditions and promotional activities, seasonal factors, holiday effects, etc. can also be considered. These factors may have a significant impact on commodity prices. In addition, in order to improve the accuracy and reliability of the data, data cleaning and preprocessing techniques can be used to remove outliers and noise data to ensure the effectiveness of subsequent analysis.

[0077] In some embodiments, step S2 includes:

[0078] S21, using dynamic time window to detect abnormal transaction time points, when it is detected that the transaction price is within the time window, When the fluctuation range exceeds 3 times the standard deviation of the historical period, the original timestamp will be compensated and corrected. The correction formula is:

[0079] ;

[0080] in, represents the transaction price at time t, represents the average price during the window period, is the dynamic compensation coefficient and , is a symbolic function;

[0081] S22. Build a cross-platform user identity mapping graph, calculate the similarity of the device fingerprint hash value and the matching degree of the consumption behavior sequence, and generate a user identity confidence score. When the confidence score is greater than 0.8, it is determined to be the same user entity;

[0082] S23. Reconstruct the spatiotemporally aligned data into a three-dimensional feature tensor, the dimensions of which include the number of users, the number of time slices, and the number of feature channels, where the feature channels at least include price volatility, consumption frequency change, and environmental variable influencing factors.

[0083] It should be noted that the present invention adopts a spatiotemporal correlation processing method when processing multi-dimensional behavioral data, and generates a spatiotemporal aligned feature tensor through abnormal transaction time point detection and cross-platform user identity mapping. Abnormal transaction time point detection refers to analyzing the fluctuation of transaction prices within a specific time window to identify transaction time points that are significantly different from historical data, thereby determining whether there is abnormal transaction behavior. Cross-platform user identity mapping is based on dual verification of device fingerprint similarity and consumer behavior consistency, correlating the behavior of the same user on different platforms to ensure the accuracy and completeness of the data. This process can effectively integrate user data from different platforms and provide a more comprehensive perspective for subsequent analysis.

[0084] Specifically, the detection of abnormal transaction time points adopts a dynamic time window method. By calculating the fluctuation range of the transaction price within the time window, when the fluctuation exceeds 3 times the standard deviation of the historical period, the original timestamp is compensated and corrected. The dynamic compensation coefficient is adjusted according to the size of the fluctuation range, ranging from 5 milliseconds to 30 milliseconds. Cross-platform user identity mapping is achieved by constructing a user identity confidence score, which is calculated based on the similarity of the device fingerprint hash value and the matching degree of the consumption behavior sequence. When the confidence score is greater than 0.8, it is determined to be the same user entity. The data after spatiotemporal alignment is reconstructed into a three-dimensional feature tensor, whose dimensions include the number of users, the number of time slices, and the number of feature channels, where the feature channels at least include price volatility, consumption frequency change, and environmental variable influencing factors.

[0085] Preferably, when detecting abnormal transaction time points, the size of the time window and the range of the dynamic compensation coefficient can be adjusted based on the transaction characteristics of different commodities. For example, for high-frequency trading commodities, the time window can be set shorter, and the dynamic compensation coefficient can be appropriately reduced to improve detection sensitivity. In cross-platform user identity mapping, in addition to verifying the consistency of device fingerprints and consumption behavior, the user's social network information can also be introduced as a supplementary verification method to further improve the accuracy of user identity recognition. In addition, when constructing feature tensors, more feature channels can be added, such as user evaluation data and browsing behavior data, to enrich the data dimension and provide more comprehensive information for subsequent analysis.

[0086] In some embodiments, the differential weighted fusion in step S3 includes:

[0087] S31. Calculate the user price sensitivity index by analyzing the user's price comparison response time in historical orders, coupon usage efficiency, and preference for alternative products. The preference for alternative products should be weighted at no less than 40%.

[0088] S32. Build a user group classification model, classifying the top 10% of users in terms of price sensitivity index as the highly sensitive group, the middle 30% as the moderately sensitive group, and the remainder as the low-sensitivity group. Different price fluctuation tolerance coefficients are set for different groups.

[0089] S33. Perform group weighted pooling processing on the feature tensor. The feature channel weights of the highly sensitive group are increased by 1.5-2 times, and the weights of the low-sensitive group are reduced to 0.7-0.9 times, thereby generating a classification feature matrix with group discrimination.

[0090] It should be noted that the present invention introduces a differentiated weighted fusion mechanism based on user price sensitivity grading when processing feature tensors. User price sensitivity grading divides users into different groups based on their response to price changes, thus providing a basis for subsequent weighted fusion. Differentiated weighted fusion assigns different weights to each feature channel in the feature tensor based on the price sensitivity of each user group, generating a classification feature matrix with group differentiation. This process can more accurately reflect the transaction behavior characteristics of different user groups, providing more accurate data support for subsequent price rationality assessments.

[0091] Specifically, the calculation of a user's price sensitivity index involves multiple dimensions, including price comparison response time, coupon usage efficiency, and the tendency to select alternative products. Price comparison response time refers to the length of time a user spends viewing the price comparison page and the number of redirects, reflecting the user's level of attention to price. Coupon usage efficiency assesses the user's response to price incentives by comparing the difference in usage time after coupon redemption with the increase in order amount. The tendency to select alternative products refers to the frequency with which users switch between products in different price ranges. When this frequency exceeds twice the average level for similar users, it is considered a high price sensitivity trait. When constructing the user group classification model, the top 10% of users in the price sensitivity index are classified as highly sensitive, the middle 30% as moderately sensitive, and the remainder as low-sensitivity. Different price fluctuation tolerance coefficients are set for different groups, and the historical transaction price fluctuation threshold for the highly sensitive group is reduced by 20%-30%.

[0092] Preferably, when calculating the user price sensitivity index, the weight distribution of each dimension can be further refined. For example, for the dimension of preference for alternative product selection, considering its key role in reflecting user price sensitivity, its weight ratio can be set to no less than 40%. When constructing a user group classification model, the division ratio of user groups can be adjusted according to the actual application scenario. For example, in some markets with fierce price competition, the proportion of highly sensitive groups can be appropriately increased to more accurately capture the behavioral characteristics of price-sensitive users. In addition, for the group weighted pooling processing of feature tensors, a dynamic adjustment mechanism can be introduced to dynamically adjust the feature channel weights of highly sensitive groups and low-sensitivity groups according to real-time market data and changes in user behavior, thereby further improving the accuracy and timeliness of the classification feature matrix.

[0093] In some embodiments, step S4 includes:

[0094] S41. Generate a user's historical benchmark price curve, using an exponential decay weighting method to integrate the user's past K transaction prices, with the transaction prices in the last 30 days accounting for no less than 60% of the weight;

[0095] S42. Calculate the dynamic price deviation index by comparing the absolute difference between the real-time transaction price and the benchmark price, and superimposing the influencing factors of the real-time supply and demand index deviation to generate a normalized deviation score. The specific formula is:

[0096] ;

[0097] in, is the base price, Indicates the deviation between the real-time supply and demand index and the historical benchmark value. is the benchmark value of the supply and demand index;

[0098] S43. Construct a price rationality scoring model. When the deviation index exceeds the tolerance threshold of the user's group, activate the price rationality degradation mechanism. Every 10-point drop in the score triggers a first-level warning.

[0099] It should be noted that when evaluating the rationality of real-time transaction prices, this invention generates a comprehensive score by calculating the dynamic price deviation index and combining it with the real-time supply and demand index and the promotion intensity parameter. The dynamic price deviation index measures the difference between the real-time transaction price and the user's historical benchmark price, while the real-time supply and demand index and the promotion intensity parameter reflect the impact of the current market environment on prices. This comprehensive scoring mechanism enables a more comprehensive assessment of transaction price rationality, effectively identifying potential discriminatory pricing behavior.

[0100] Specifically, the user's historical benchmark price curve is generated by fusing the user's past K transaction prices through the exponential decay weighting method, where the weight of the transaction price in the last 30 days is no less than 60%, to ensure that the benchmark price can reflect the user's recent consumption. The dynamic price deviation index is generated by comparing the absolute difference between the real-time transaction price and the benchmark price, and superimposing the influencing factor of the real-time supply and demand index deviation. The real-time supply and demand index deviation is obtained by collecting real-time market data, including the median price of competing platforms, inventory turnover rate, search popularity index, etc., calculating the supply and demand balance coefficient and normalizing it. The price rationality scoring model activates the price rationality degradation mechanism based on whether the deviation index exceeds the tolerance threshold of the user's group, and triggers an early warning when the score is lower than the preset risk level.

[0101] Preferably, when generating the user's historical benchmark price curve, the weight distribution of the past K transaction prices can be dynamically adjusted according to the consumption cycle and market volatility characteristics of different commodities. For example, for seasonal commodities, the weight of the transaction price in the past quarter can be increased, while for daily consumer goods, more emphasis can be placed on the transaction price in the past month. When calculating the dynamic price deviation index, more market factors can be introduced, such as the type and intensity of promotional activities, to further refine the calculation method of the deviation degree of the real-time supply and demand index. For example, for limited-time discounts and full-reduction activities, different influencing factor weights can be set respectively to more accurately reflect the impact of promotional activities on price rationality. In addition, in the price rationality scoring model, the tolerance threshold can be dynamically adjusted according to the price sensitivity of the user group. For user groups with high price sensitivity, stricter thresholds can be set to improve the accuracy of the early warning.

[0102] In some embodiments, the legal terms matching engine workflow in step S5 includes:

[0103] S51. Construct a legal knowledge graph, parse the relevant anti-price discrimination laws into structured (illegal behavior, identification requirements, penalty measures) triples, and map them into a high-dimensional feature space;

[0104] S52. Calculate the matching degree between the real-time scenario feature vector and the legal provision feature vector, focusing on the three core dimensions of price difference rationality, user identity relevance, and subjective intent determination, and set a dynamic weight coefficient for each dimension;

[0105] S53. Generate a legal basis report. When the matching degree exceeds 0.75, it will automatically link the precedents of similar cases in the past three years, and mark the median penalty amount and common defense reasons.

[0106] It should be noted that the present invention uses a legal clause matching engine to match warning scenario feature vectors with legal provisions, generating a warning report containing the legal basis and penalty clauses for anti-price discrimination. The legal clause matching engine utilizes a legal knowledge graph and similarity calculations between scenario feature vectors to implement a legal assessment of differential pricing behavior. The legal knowledge graph parses anti-price discrimination-related legal provisions into structured triples (illegal behavior, identification requirements, and penalty measures) and maps them into a high-dimensional feature space for matching with real-time scenario feature vectors. This mechanism can provide regulators and users with a clear legal basis, enhancing the deterrent effect against differential pricing behavior.

[0107] Specifically, the workflow of the legal clause matching engine includes three main steps. First, when constructing the legal knowledge graph, natural language processing technology is used to parse the legal text, extract the key elements of the constituent elements of the illegal behavior, such as the subject of the behavior, subjective aspects, objective behavior and damage results, and map these elements to a 64-dimensional semantic vector space. Secondly, the matching degree between the real-time scene feature vector and the legal clause feature vector is calculated, focusing on matching the three core dimensions of price difference rationality, user identity relevance, and subjective intention determination, and setting a dynamic weight coefficient for each dimension. Finally, when the matching degree exceeds 0.75, a legal basis report is automatically generated, and the precedents of similar cases in the past three years are linked, marking the median penalty amount and common defense reasons. This process ensures the accuracy and timeliness of legal evaluation.

[0108] Preferably, when constructing a legal knowledge graph, the parsing process of legal provisions can be further refined. For example, for penalty amount ranges, a discretized coding method can be used to convert continuous amount ranges into discrete categories to facilitate subsequent matching and analysis. At the same time, for the statute of limitations, it can be converted into an attenuation coefficient to reflect the impact of time on legal effectiveness. When calculating the matching degree, the weight coefficients of the three core dimensions can be dynamically adjusted according to the complexity of different scenarios. For example, in transactions involving high-value goods, the weight of the price difference rationality dimension can be increased, while in scenarios involving user privacy data, the weight of the user identity relevance dimension can be increased. In addition, to improve the practicality of the legal basis report, a detailed explanation of the relevant legal provisions and an explanation of the applicable conditions can be added to the report to help users and regulators better understand and apply the legal basis.

[0109] In some embodiments, the method for calculating the user price sensitivity index further includes:

[0110] S311. Analyze the user's price comparison behavior characteristics, count the user's stay time and number of jumps to the price comparison page, and generate a price attention index;

[0111] S312. Calculate the coupon usage efficiency by comparing the difference in usage time after coupon redemption with the increase in order amount to assess the user's response speed to price incentives.

[0112] S313. Build an alternative product selection model to record the frequency of users switching between products in different price ranges. When the frequency is more than twice the average level of similar users, it is determined to be a high price sensitivity feature.

[0113] It should be noted that the present invention further refines the analysis methods of user price comparison behavior characteristics, coupon usage efficiency and alternative product selection tendency when calculating the user price sensitivity index. User price comparison behavior characteristics refer to the behavior pattern of users viewing price comparison pages during the shopping process, including the length of stay and the number of jumps, so as to measure the user's attention to price. Coupon usage efficiency is to evaluate the user's response speed to price incentives by analyzing the user's usage after receiving the coupon. Alternative product selection tendency refers to the frequency of users switching products within different price ranges. This indicator can effectively reflect the user's sensitivity to price changes. By comprehensively evaluating these three dimensions, the user's price sensitivity index can be calculated more accurately.

[0114] Specifically, analysis of user price comparison behavior includes measuring the duration of time users spend viewing the price comparison page and the number of times they navigate to other product pages, generating a price attention index. For example, if a user spends a long time on the price comparison page and frequently navigates to other product pages, this indicates that the user is price-sensitive, and their price attention index will increase accordingly. Coupon usage efficiency is evaluated by comparing the difference in usage time after coupon redemption with the increase in order value. For example, if a user uses a coupon within a short period of time after receiving it and the order value increases significantly, this indicates that the user is responsive to price incentives and has high coupon usage efficiency. Analysis of product substitution preferences records the frequency with which users switch between products in different price ranges. When this frequency exceeds twice the average for similar users, it is considered a high price sensitivity trait. For example, if a user frequently switches between different brands or models of products within the same price range, this indicates that they are sensitive to price changes.

[0115] Preferably, when analyzing the characteristics of user price comparison behavior, the weight settings of the dwell time and the number of jumps can be further refined. For example, the weight ratio of the dwell time and the number of jumps can be dynamically adjusted according to different product categories and market environments. For high-value products, the weight of the dwell time can be increased because users usually spend more time comparing when purchasing such products. When calculating the efficiency of coupon usage, more parameters can be introduced, such as the denomination of the coupon, the usage scenario, etc., to more comprehensively evaluate the user's response to price incentives. For example, for large coupons, users may be more inclined to use them on high-value products, so different usage efficiency evaluation criteria can be set for coupons of different denominations.

[0116] Furthermore, when analyzing the tendency to choose alternative products, the user purchase conversion rate can be introduced as an auxiliary indicator to more accurately reflect the user's price sensitivity. For example, if a user frequently switches products before finally completing a purchase, it indicates that they are highly sensitive to price changes, but a high purchase conversion rate also indicates that the user is looking for products with better value for money.

[0117] In some embodiments, the method for calculating the supply-demand index deviation includes:

[0118] S421. Collect real-time market data, including median prices on competing platforms, inventory turnover rates, and search popularity indexes;

[0119] S422. Calculate the supply-demand balance coefficient:

[0120] ;

[0121] in, The average price of competing products refers to the median price of other platforms in the same or similar product categories. The real-time search volume, that is, the search volume of the current product, reflects the market demand for the product. The current inventory, representing the current inventory quantity of the product, reflects the market supply situation. is the supply-demand balance coefficient, The mean search volume is the average of historical search volumes and is used to calculate relative changes in search volume to measure fluctuations in market demand. For the price of products on this platform, The average inventory level is the historical average inventory level, which is used as a comparison benchmark to assess the difference between the current inventory level and the historical average level;

[0122] S423. Normalize the supply-demand balance coefficient to generate an impact factor in the range [0,1].

[0123] It should be noted that when calculating the supply and demand index deviation, the present invention generates an impact factor by collecting real-time market data and calculating the supply and demand balance coefficient. The supply and demand index deviation reflects the difference between the current market supply and demand relationship and the historical average level, and is an important reference indicator for evaluating price rationality. Real-time market data includes the median price of competing platforms, inventory turnover rate, and search popularity index, etc. These data can fully reflect market dynamics. The supply and demand balance coefficient is obtained by weighted calculation of these data, and the final generated normalized impact factor is used to adjust the relevant parameters in the price rationality scoring model.

[0124] Specifically, when calculating the supply and demand index deviation, real-time market data is first collected, including the median price of competing platforms, inventory turnover rate, and search popularity index. The median price of competing platforms refers to the median price of other platforms in the same or similar product categories; inventory turnover rate refers to the ratio of current inventory to historical average inventory; and search popularity index refers to the ratio of the current product search volume to the historical average search volume. The supply and demand balance coefficient is obtained by weighted calculation of these data, where the weight of each parameter can be adjusted according to the market characteristics and historical data of the product. For example, for seasonal products, the weight of the search popularity index can be appropriately increased because it is more affected by seasonality. Finally, the supply and demand balance coefficient is normalized to generate an impact factor in the interval [0,1], which is used to adjust the relevant parameters in the price rationality scoring model.

[0125] Preferably, when collecting real-time market data, the data source and collection frequency can be further refined. For example, for the median price of competing platforms, the frequency of data collection can be increased from once an hour to once every 15 minutes to reflect market dynamics more promptly. When calculating the supply and demand balance coefficient, the weights of each parameter can be dynamically adjusted according to the market volatility characteristics of different commodities. For example, for commodities with high demand and low inventory, the weight of inventory turnover rate can be increased to more accurately reflect the market supply and demand relationship. In addition, different methods can be used for normalization, such as maximum and minimum normalization or Z-score normalization. The specific method can be selected according to the distribution characteristics of the data. For example, when the data distribution is relatively uniform, maximum and minimum normalization can be used; when there are outliers in the data distribution, Z-score normalization can be used to improve the accuracy and stability of the influencing factors.

[0126] In some embodiments, the method for constructing the legal text feature vector includes:

[0127] S511. Use natural language processing technology to parse legal texts and extract key elements of the elements of illegal behavior, including the subject of the behavior, subjective aspects, objective behavior, and damage results;

[0128] S512, mapping each element to a 64-dimensional semantic vector space, and calculating the association weights between elements through the attention mechanism;

[0129] S513. When generating the feature vector of the article, the penalty amount interval is discretized and encoded, and the limitation period is converted into an attenuation coefficient.

[0130] It should be noted that when constructing legal text feature vectors, the present invention uses natural language processing technology to parse legal texts, extract key elements that constitute the elements of an illegal act, and map these elements into a high-dimensional semantic vector space. The construction of legal text feature vectors is the core component of the legal clause matching engine. By converting the semantic information of legal texts into a computable vector form, it provides the foundation for subsequent matching and evaluation. This process not only improves the operability of legal texts but also enhances the accuracy and efficiency of legal evaluation.

[0131] Specifically, when constructing a feature vector for a legal provision, natural language processing technology is first used to parse the legal provision and extract the key elements of the elements of an illegal act, including the subject, subjective aspects, objective behavior, and damages. The subject refers to the individual or organization that commits the illegal act; the subjective aspect refers to the subjective intent of the perpetrator, such as intent or negligence; the objective behavior refers to the specific actions taken, such as price discrimination; and the damages refer to the specific consequences of the act, such as damage to consumer rights. These elements are mapped into a 64-dimensional semantic vector space, and the correlation weights between them are calculated using an attention mechanism. When generating the feature vector for the provision, the penalty amount range is discretized and the statute of limitations is converted into a decay coefficient. Discretization divides the continuous penalty amount range into discrete categories to facilitate subsequent matching and analysis. The decay coefficient quantifies the timeliness of the legal provision based on the length of the statute of limitations.

[0132] Preferably, when extracting the key elements of legal texts, the application of natural language processing technology can be further refined. For example, pre-trained language models in deep learning (such as BERT or Transformer) are used to improve the accuracy and efficiency of text parsing. When mapping to the semantic vector space, the dimension of the vector space can be dynamically adjusted according to the complexity of the legal text. For example, for complex texts involving multiple legal fields, the dimension of the vector space can be increased to better capture semantic information. When discretizing the encoding of penalty amount ranges, the amount ranges can be flexibly divided according to the actual circumstances of different legal fields or specific cases. For example, in cases involving high-value goods, the penalty amount range can be divided into finer categories,

[0133] In addition, when converting the limitation period into a decay coefficient, a time decay function, such as exponential decay or logarithmic decay, can be introduced to more accurately reflect the impact of time on legal effectiveness.

[0134] In some embodiments, the multi-level intervention strategy in step S6 includes:

[0135] S61: At the high-risk warning level, execute simultaneously:

[0136] Real-time freezing of the price adjustment interface for suspected abnormal products

[0137] Show users the price comparison information flow of the product on different user terminals

[0138] Automatically generate electronic evidence packages containing timestamps, price comparison charts, and legal basis;

[0139] S62, medium-risk warning level, triggers:

[0140] Extend the price update review period to a minimum of 2 hours

[0141] Send a second confirmation request to the operator, and provide a justification for the price adjustment;

[0142] S63, low-risk warning level, execute:

[0143] Record abnormal transaction logs and mark suspected feature labels

[0144] Start tracking and monitoring of similar transaction behaviors for the next 7 days.

[0145] It should be noted that when implementing the multi-level intervention strategy, the present invention adopts different intervention measures according to the level of the early warning report. The multi-level intervention strategy aims to take corresponding measures to deal with potential differential pricing behaviors through early warning signals at different levels. At the high-risk warning level, measures such as freezing the price adjustment interface, displaying the price comparison information flow, and generating an electronic evidence package will be executed simultaneously; at the medium-risk warning level, it will trigger the extension of the price update review cycle and the sending of a secondary confirmation request; at the low-risk warning level, it will record abnormal transaction logs and start tracking monitoring. This hierarchical intervention mechanism can effectively balance regulatory intensity and market flexibility, ensuring that while protecting consumer rights, it does not affect the normal operation of the market.

[0146] Specifically, the implementation of a multi-tiered intervention strategy involves several specific operational steps. At the high-risk warning level, the system will immediately freeze the price adjustment interface for suspected abnormal products, preventing merchants from further adjusting prices. Simultaneously, users will be presented with a flow of price comparisons for the product across different user terminals, helping them understand price differences. Furthermore, the system will automatically generate an electronic forensics package containing timestamps, price comparison charts, and legal basis to provide supporting evidence when needed. At the medium-risk warning level, the system will extend the price update review cycle to a minimum of two hours, providing regulators with ample time to review the rationale for price adjustments. Simultaneously, a second confirmation request will be sent to operators, requesting an explanation for the price adjustment. At the low-risk warning level, the system will log abnormal transactions and label them with suspected characteristics. It will also initiate tracking and monitoring of similar transactions for the next seven days to continuously monitor market dynamics.

[0147] Preferably, when implementing a multi-tiered intervention strategy, the specific actions for each warning level can be further refined. For example, at the high-risk warning level, in addition to the aforementioned measures, real-time warnings can be added to merchants, alerting them to the potential legal risks of price adjustments. At the medium-risk warning level, an AI-assisted review mechanism can be introduced to quickly analyze the justifications provided by operators, improving review efficiency. At the low-risk warning level, an automatic flagging function for abnormal transactions can be added to facilitate rapid problem identification during subsequent analysis. Furthermore, the warning level thresholds can be dynamically adjusted based on market dynamics and user feedback. For example, during periods of significant market volatility, the warning threshold can be appropriately raised to reduce false positives; during periods of market stability, the threshold can be lowered to increase warning sensitivity.

[0148] In some embodiments, a model iterative optimization mechanism is also included:

[0149] S71. Collect user feedback data and subsequent transaction behavior data after the warning is triggered, and construct a feedback training set;

[0150] S72. Optimize the parameters of the price rationality scoring model, focusing on adjusting the proportional relationship between the supply and demand index influencing factors and the user group weight coefficients. The optimization objective function is:

[0151] ;

[0152] in, 、 is the balance coefficient, which is used to adjust the weights of the mean absolute error (MAE) and KL divergence in the optimization objective function to achieve a balance between the accuracy and generalization ability of the model;

[0153] Represents the mean absolute error, which is used to measure the average absolute difference between the model prediction value and the true value;

[0154] The KL divergence of the warning distribution is used to measure the similarity between the model prediction distribution P and the true distribution Q in the optimization objective function. The smaller the KL divergence value, the more similar the two distributions are. That is, the closer the model prediction distribution is to the true distribution, the stronger the generalization ability of the model.

[0155] The model prediction value is the prediction result of the price rationality scoring model on the transaction price rationality score. By optimizing the model parameters, A price-reasonability score as close to reality as possible ;

[0156] The true value refers to the actual price rationality score. It is the benchmark value used for comparison during model training and optimization, reflecting the rationality of the transaction price in actual conditions.

[0157] The model prediction distribution is the probability distribution of the prediction results of the price rationality scoring model on the transaction price rationality score, reflecting the model's prediction possibility for different scoring situations;

[0158] The real distribution refers to the actual probability distribution of the transaction price rationality score, which represents the real situation of the transaction price rationality score in the market. The model optimizes the parameters to make the predicted distribution Try to get as close to the true distribution as possible .

[0159] It should be noted that the present invention incorporates a model iteration optimization mechanism. By collecting user feedback data and subsequent transaction behavior data after the early warning is triggered, a feedback training set is constructed to optimize the parameters of the price rationality scoring model. This mechanism aims to improve the model's accuracy in identifying differential pricing behavior and the reliability of early warnings through continuous learning and optimization. The core of the model iteration optimization mechanism lies in dynamically adjusting the proportional relationship between the supply and demand index influencing factors and the user group weight coefficients to better adapt to market changes and the dynamic characteristics of user behavior.

[0160] Specifically, the model iterative optimization mechanism involves two main steps. First, user feedback data is collected after the warning is triggered. This data includes the user's response to the warning information, subsequent trading behavior, and evaluation of the price adjustment. This data is used to construct a feedback training set, providing a basis for model optimization. Second, the parameters of the price rationality scoring model are optimized, focusing on adjusting the proportional relationship between the supply and demand index influencing factors and the user group weight coefficient. The optimization objective function is achieved by minimizing the weighted sum of the mean absolute error (MAE) and the KL divergence. The mean absolute error measures the difference between the model's predicted value and the actual value, while the KL divergence is used to measure the similarity between the model's predicted distribution and the true distribution. By adjusting the balance coefficient, a balance can be achieved between the model's accuracy and generalization ability.

[0161] Preferably, when constructing the feedback training set, the method of collecting user feedback data can be further refined. For example, user satisfaction with warning information and subsequent behavior can be collected through user surveys, online feedback forms, or feedback functions in mobile applications. When optimizing the price rationality scoring model, the supply and demand index influencing factors and user group weight coefficients can be dynamically adjusted according to the characteristics of different market environments and user groups. For example, in markets with large demand fluctuations, the weight of the supply and demand index influencing factors can be increased; in markets with large differences in user groups, the weight of the user group weight coefficient can be increased. In addition, more optimization algorithms, such as genetic algorithms or Bayesian optimization, can be introduced to improve the efficiency and effectiveness of model parameter optimization. These algorithms can automatically adjust parameters according to the performance of the model, further improving the accuracy and reliability of the model.

[0162] The above-mentioned embodiments of the present invention have the following beneficial effects: the present invention can realize accurate identification and real-time early warning of abnormal price behavior. By collecting multi-dimensional behavior data of users in real time and performing spatiotemporal correlation processing, combined with differentiated weighted fusion of user price sensitivity grading, it is possible to accurately capture abnormal transaction characteristics and generate a classification feature matrix with high discrimination. On this basis, the dynamic deviation index is calculated and a comprehensive score of price rationality is generated. When the score is lower than the preset risk level, the early warning mechanism is triggered, thereby timely discovering potential abnormal price behavior, providing users with a more transparent consumption environment, helping them make more informed consumption decisions, and also providing technical support for enterprises to assist them in optimizing pricing strategies.

[0163] Furthermore, the present invention can also provide legal basis and relevant clause prompts for early warning scenarios. The early warning scenario feature vector is input into the legal clause matching engine. Through semantic analysis of legal text and calculation of similarity between the scenario feature vectors, an early warning report containing the relevant legal clause basis and reasonable explanation is generated. This not only helps users protect their legitimate rights and interests when encountering abnormal pricing behavior, but also provides clear compliance guidance for enterprises, improves their operational efficiency, and promotes the rationality and fairness of market pricing.

[0164] The present invention can also effectively curb the occurrence of abnormal pricing behavior through a multi-level intervention strategy, taking corresponding measures according to different warning levels. At the high-risk warning level, the price adjustment interface of suspected abnormal products can be frozen in real time, showing the user the price comparison information flow of the product at different user terminals, and automatically generating an electronic evidence package containing a timestamp, price comparison chart, and legal basis; at the medium-risk warning level, the price update review cycle can be extended to a minimum of 2 hours, and a second confirmation request can be sent to the operator, requiring an explanation of the rationality of the price adjustment; at the low-risk warning level, abnormal transaction logs can be recorded and marked with suspected feature tags, and tracking and monitoring of similar transactions for the next 7 days can be initiated.

[0165] In addition, the present invention can also collect user feedback data and subsequent transaction behavior data after the early warning is triggered through a model iteration optimization mechanism, construct a feedback training set, optimize the price rationality scoring model parameters, and focus on adjusting the proportional relationship between the supply and demand index influencing factors and the user group weight coefficients to further improve the accuracy and reliability of the model.

[0166] Furthermore, the storage medium of the embodiment of the present application stores program instructions that can implement all the above methods, wherein the program instructions can be stored in the above storage medium in the form of a software product, including a number of instructions for causing a computer device (which can be a personal computer, server, or network device, etc.) or a processor to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, or a terminal device such as a computer, server, mobile phone, or tablet.

[0167] The above descriptions merely illustrate some preferred embodiments of the present invention and the underlying technical principles. Those skilled in the art should understand that the scope of the invention encompassed by the embodiments of the present invention is not limited to technical solutions formed by specific combinations of the aforementioned technical features. It also encompasses other technical solutions formed by any combination of the aforementioned technical features or their equivalents, without departing from the aforementioned inventive concept. For example, a technical solution formed by replacing the aforementioned features with (but not limited to) technical features with similar functions disclosed in the embodiments of the present invention.

Claims

1. A method for identifying, warning and legal reminder based on big data, characterized in that: The following steps are involved: S1. Real-time collection of multi-dimensional user behavior data, including user identity feature data, historical transaction price series, and real-time environmental variable data. The user identity feature data includes at least device fingerprint features, membership level tags, and spending capacity assessment values, which are calculated by weighting the user's historical order amount and consumption frequency. S2. Performing spatiotemporal correlation processing on the multi-dimensional behavioral data to generate a spatiotemporally aligned feature tensor by detecting abnormal transaction time points and mapping user identities across platforms, wherein the cross-platform user identity mapping is based on dual verification of device fingerprint similarity and consumer behavior consistency; S3. Based on the user price sensitivity classification results, perform differentiated weighted fusion on the feature tensor to generate a user group classification feature matrix, wherein the historical transaction price fluctuation threshold of the high price sensitivity user group is reduced by 20%-30%. The differentiated weighted fusion includes: S31. Calculate the user price sensitivity index by analyzing the price comparison response time, coupon usage efficiency, and preference for alternative products in the user's historical orders to conduct a comprehensive assessment. S32. Build a user group classification model, classifying the top 10% of users in terms of price sensitivity index as the highly sensitive group, the middle 30% as the moderately sensitive group, and the remainder as the low-sensitivity group. Different price fluctuation tolerance coefficients are set for different groups. S33, performing group weighted pooling processing on the feature tensor, increasing the feature channel weights of the highly sensitive group, and generating a classification feature matrix with group discrimination; S4. Calculate the dynamic deviation index of the real-time transaction price relative to the user's historical benchmark price, and integrate the real-time supply and demand index with the promotion intensity parameter to generate a comprehensive score for price rationality. When the score falls below the preset risk level, trigger an early warning mechanism; S5. Input the warning scenario feature vector into a legal clause matching engine to generate a warning report containing the anti-price discrimination legal basis and penalty clauses. The legal clause matching engine is implemented by semantic analysis of legal provisions and similarity calculation of scenario feature vectors. S6. Implement multi-level intervention strategies based on the warning report level, including pushing real-time price comparison information to user terminals, sending interface flow limiting instructions to the enterprise risk control system, and submitting risk filing data packages to the regulatory platform.

2. The method according to claim 1, characterized in that Step S2 includes: S21. Use a dynamic time window to detect abnormal transaction time points. When it is detected that the fluctuation range of the transaction price within the time window [tw, t] exceeds 3 times the standard deviation of the historical period, the original timestamp is compensated and corrected. The correction formula is: Among them, p t represents the transaction price at time t, represents the average price during the window period, δ is the dynamic compensation coefficient, sgn() is the sign function, t correct is the corrected timestamp, t raw is the original timestamp; S22. Build a cross-platform user identity mapping graph, calculate the similarity of the device fingerprint hash value and the matching degree of the consumption behavior sequence, and generate a user identity confidence score. When the confidence score is greater than 0.8, it is determined to be the same user entity; S23. Reconstruct the spatiotemporally aligned data into a three-dimensional feature tensor, the dimensions of which include the number of users, the number of time slices, and the number of feature channels, where the feature channels at least include price volatility, consumption frequency change, and environmental variable influencing factors.

3. The method according to claim 1, characterized in that Step S4 includes: S41. Generate a user's historical benchmark price curve, using an exponential decay weighting method to integrate the user's past K transaction prices, with the weight of the transaction prices in the last 30 days accounting for no less than 60%; S42. Calculate the dynamic price deviation index by comparing the absolute difference between the real-time transaction price and the benchmark price, and superimposing the influencing factors of the real-time supply and demand index deviation to generate a normalized deviation score. The specific formula is: in, is the benchmark price, ΔE represents the deviation between the real-time supply and demand index and the historical benchmark value, E base is the benchmark value of the supply and demand index, D t is the normalized deviation score, p t This is the real-time transaction price; S43. Construct a price rationality scoring model. When the deviation index exceeds the tolerance threshold of the user's group, activate the price rationality degradation mechanism. Every 10-point drop in the score triggers a first-level warning.

4. The method according to claim 1, wherein The legal terms matching engine workflow in step S5 includes: S51. Construct a legal knowledge graph to parse anti-price discrimination related legal provisions into structured triplets of illegal behavior, identification requirements, and penalty measures, and map them into a high-dimensional feature space; S52. Calculate the matching degree between the real-time scenario feature vector and the legal provision feature vector, focusing on the three core dimensions of price difference rationality, user identity relevance, and subjective intent determination, and set a dynamic weight coefficient for each dimension; S53. Generate a legal basis report. When the matching degree exceeds 0.75, it will automatically link the precedents of similar cases in the past three years, and mark the median penalty amount and common defense reasons.

5. The method according to claim 3, characterized in that The method for calculating the user price sensitivity index also includes: S311. Analyze the user's price comparison behavior characteristics, count the user's stay time and number of jumps to the price comparison page, and generate a price attention index; S312. Calculate the coupon usage efficiency by comparing the difference in usage time after coupon redemption with the increase in order amount to assess the user's response speed to price incentives. S313. Build an alternative product selection model to record the frequency of users switching between products in different price ranges. When the frequency is more than twice the average level of similar users, it is determined to be a high price sensitivity feature.

6. The method according to claim 3, characterized in that The method for calculating the deviation between the real-time supply and demand index and the historical benchmark value includes: S421. Collect real-time market data, including median prices on competing platforms, inventory turnover rates, and search popularity indexes; S422. Calculate the supply-demand balance coefficient: Among them, P compete is the average price of competing products, S search is the real-time search volume, I stock is the current inventory, ΔE is the supply and demand balance coefficient, is the mean search volume, P ours For the price of goods on this platform, I avg is the average inventory level; S423. Normalize the supply-demand balance coefficient to generate an impact factor in the range [0,1].

7. The method according to claim 4, characterized in that The method for constructing the legal text feature vector includes: S511. Use natural language processing technology to parse legal texts and extract key elements of the elements of illegal behavior, including the subject of the behavior, subjective aspects, objective behavior, and damage results; S512, mapping each element to a 64-dimensional semantic vector space, and calculating the association weights between elements through the attention mechanism; S513. When generating the feature vector of the article, the penalty amount interval is discretized and encoded, and the limitation period is converted into an attenuation coefficient.

8. The method according to claim 1, characterized in that The multi-level intervention strategy in step S6 includes: S61: At the high-risk warning level, execute simultaneously: Real-time freezing of the price adjustment interface for suspected abnormal products; Show users the price comparison information flow of the product on different user terminals; Automatically generate electronic evidence packages containing timestamps, price comparison charts, and legal basis; S62, medium-risk warning level, triggers: Extend the price update review period to a minimum of 2 hours; Send a second confirmation request to the operator, and provide a justification for the price adjustment; S63, low-risk warning level, execute: Record abnormal transaction logs and mark them with suspected feature labels; Start tracking and monitoring of similar transaction behaviors for the next 7 days.

9. The method according to claim 3, characterized in that It also includes a model iterative optimization mechanism: S71. Collect user feedback data and subsequent transaction behavior data after the warning is triggered, and construct a feedback training set; S72. Optimize the parameters of the price rationality scoring model, focusing on adjusting the proportional relationship between the supply and demand index influencing factors and the user group weight coefficients. The optimization objective function is: L=α·MAE(y pred ,y true )+β·KL(P‖Q); Among them, α and β are balance coefficients, MAE represents mean absolute error, KL(P‖Q) is the KL divergence of the warning distribution, and y pred is the model prediction value, y true is the true value, P is the model prediction distribution, and Q is the true distribution.

Images