Behavior data-based bidding auction system anti-cheating method and device

Abstract

The invention provides a bidding auction system anti-cheating method and device based on behavior data, and relates to the technical field of data processing, and the method comprises the steps: obtaining the behavior data of a target supplier in a bidding auction system; inputting the behavior data into a pre-trained behavior recognition model to obtain a behavior analysis result; wherein the behavior recognition model is constructed through a machine learning algorithm based on historical behavior data; and determining a risk level of the target supplier according to the behavior analysis result and a preset rule, and executing a risk control measure corresponding to the risk level. Through the method provided by the invention, the behavior recognition model is constructed, whether the quotation behavior of the supplier is manual operation can be accurately recognized, the supplier is effectively prevented from automatically quoting through a third-party cheating tool, and fairness and fairness of bidding auction are ensured. Meanwhile, through real-time monitoring and automatic judgment, manual intervention is reduced, and the anti-cheating efficiency and accuracy are improved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to an anti-cheating method and device for a bidding auction system based on behavior data. Background Art

[0002] With the rapid development of railway informatization, railway bidding systems have gradually become a key method for railway material procurement. However, some suppliers use third-party cheating tools to automatically bid, seriously affecting the fairness of bidding auctions. Existing anti-cheating methods mainly rely on measures such as IP restrictions and verification codes, but these methods are easily cracked and cannot effectively detect cheating behavior in automated tools.

[0003] How to accurately identify and prevent automated quotation cheating is a technical problem that needs to be solved at present. Summary of the Invention

[0004] The present invention provides a method and device for preventing cheating in a bidding auction system based on behavioral data, so as to solve the defects in the prior art.

[0005] The present invention provides a method for preventing cheating in a bidding auction system based on behavioral data, comprising the following steps: Obtaining the target supplier's behavior data in the bidding auction system; Inputting the behavior data into a pre-trained behavior recognition model to obtain a behavior analysis result; wherein the behavior recognition model is constructed based on historical behavior data through a machine learning algorithm; The risk level of the target supplier is determined based on the behavior analysis results and preset rules, and risk control measures corresponding to the risk level are executed.

[0006] According to a method for preventing cheating in a bidding auction system based on behavioral data provided by the present invention, obtaining behavioral data of a target supplier in the bidding auction system includes: Real-time collection of initial behavior data of the target supplier in the bidding auction system; wherein the initial behavior data includes: bidding operation frequency, bidding time interval, mouse sliding trajectory, keyboard input mode and page dwell time; The initial behavior data is cleaned and standardized, and key features are extracted to obtain the behavior data.

[0007] According to the present invention, a method for preventing cheating in a bidding auction system based on behavioral data is provided. The process of constructing the behavior recognition model includes: Obtaining historical behavior data of multiple suppliers in a bidding auction system and extracting historical key features of the historical behavior data; wherein the historical behavior data includes: normal behavior data and cheating behavior data of each of the multiple suppliers; Filtering the target historical key features from the historical key features using a machine learning algorithm to determine the target historical key features; wherein the target historical key features are: features that have the greatest impact on behavior recognition; Building a baseline model based on the target's historical key features; wherein the baseline model is used to: define a normal behavior pattern; The behavior recognition model is constructed based on the baseline model and the target historical key features; wherein the behavior recognition model is used to: analyze the behavior data of the target supplier in real time and identify abnormal behavior patterns.

[0008] According to a method for preventing cheating in a bidding auction system based on behavioral data provided by the present invention, the risk levels include: pass, warning, and rejection; Determining the risk level of the target supplier based on the behavior analysis results and preset rules, and executing risk control measures corresponding to the risk level, includes: If it is determined that the risk level of the target supplier is passing, executing a first risk control measure corresponding to the passing risk level; wherein the first risk control measure includes: allowing quotation, continuous monitoring, and experience optimization; When it is determined that the risk level of the target supplier is a warning, a second risk control measure corresponding to the risk level of the warning is executed; wherein the second risk control measure includes: graphic verification; When it is determined that the risk level of the target supplier is rejection, a third risk control measure corresponding to the risk level of rejection is executed; wherein, the third risk control measure includes: prohibiting quotation, cheating prompt, recording violation behavior and alarm notification.

[0009] According to a method for preventing cheating in a bidding auction system based on behavioral data provided by the present invention, when determining that the risk level of the target supplier is a warning, executing a second risk control measure corresponding to the risk level of the warning includes: If the risk level of the target supplier is determined to be a warning, a graphic verification code is sent to the target supplier for verification; If the target supplier passes the verification within a preset number of times, executing the first risk control measure corresponding to the risk level of passing; In the event that the target supplier fails verification within a preset number of times, the third risk control measure corresponding to the risk level of rejection is executed.

[0010] According to a method for preventing cheating in a bidding auction system based on behavioral data provided by the present invention, after determining the risk level of the target supplier based on the behavioral analysis results and preset rules and executing risk control measures corresponding to the risk level, the method further includes: Optimizing the behavior recognition model based on the first feedback result and the second feedback result; The first feedback result is obtained by regularly evaluating the risk control measures; and the second feedback result is obtained by analyzing the behavior analysis results to obtain the target cheating method.

[0011] According to a method for preventing cheating in a bidding auction system based on behavioral data provided by the present invention, after determining that the risk level of the target supplier is rejected and executing a third risk control measure corresponding to the risk level of rejected, the method further includes: Obtaining the appeal result of the target supplier with a risk level of rejected, and re-evaluating based on the appeal result to obtain an evaluation result; The risk control measures are adjusted based on the appeal result and the assessment result.

[0012] The present invention also provides an anti-cheating device for a bidding auction system based on behavioral data, comprising the following modules: A behavior data acquisition module is used to obtain the behavior data of the target supplier in the bidding auction system; A behavior analysis module, configured to input the behavior data into a pre-trained behavior recognition model to obtain behavior analysis results; wherein the behavior recognition model is constructed using a machine learning algorithm based on historical behavior data; The risk identification and control module is used to determine the risk level of the target supplier based on the behavior analysis results and preset rules, and to implement risk control measures corresponding to the risk level.

[0013] The present invention also provides an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the anti-cheating method for a bidding auction system based on behavioral data as described above is implemented.

[0014] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the anti-cheating method for a bidding auction system based on behavioral data as described above is implemented.

[0015] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-mentioned anti-cheating methods for a bidding auction system based on behavioral data.

[0016] The present invention provides a method and device for preventing cheating in a bidding auction system based on behavioral data, which obtains the behavioral data of a target supplier in the bidding auction system; inputs the behavioral data into a pre-trained behavior recognition model to obtain a behavioral analysis result; wherein the behavior recognition model is constructed based on historical behavioral data through a machine learning algorithm; determines the risk level of the target supplier according to the behavioral analysis result and preset rules, and executes risk control measures corresponding to the risk level. It can be seen that the present invention monitors the behavioral data of suppliers in the bidding auction system and constructs a behavior recognition model, which can accurately identify whether the supplier's bidding behavior is a human operation, effectively prevent suppliers from automatically quoting through third-party cheating tools, and ensure the fairness and impartiality of the bidding auction. At the same time, the present invention reduces manual intervention and improves the efficiency and accuracy of anti-cheating through real-time monitoring and automated judgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, a brief introduction is given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0018] Figure 1 It is a flow chart of the anti-cheating method for a bidding auction system based on behavioral data provided by the present invention.

[0019] Figure 2 This is a complete flow chart of the anti-cheating method for a bidding auction system based on behavioral data provided by the present invention.

[0020] Figure 3 It is a structural diagram of the anti-cheating device for the bidding auction system based on behavioral data provided by the present invention.

[0021] Figure 4 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0022] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only some of the embodiments of the present invention, not all of them. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0023] The following combination Figures 1-4The present invention describes a method and device for preventing cheating in a bidding auction system based on behavioral data.

[0024] Figure 1 This is a flow chart of the anti-cheating method for a bidding auction system based on behavioral data provided by the present invention. Figure 1 As shown, the method includes the following: Step 100: Obtain the target supplier's behavior data in the bidding auction system.

[0025] Figure 2 This is a complete flow chart of the anti-cheating method for the bidding auction system based on behavioral data provided by the present invention. Figure 2 , the anti-cheating method for the bidding auction system based on behavioral data provided by the present invention is specifically described.

[0026] It should be noted that this embodiment uses front-end technology (such as JavaScript) to collect supplier behavior data in real time and transmits it to the back-end through WebSocket or long polling technology.

[0027] Step 100 obtains the target supplier's behavior data in the bidding auction system, including: Step 110: Collect the initial behavior data of the target supplier in the bidding auction system in real time; wherein, the initial behavior data includes: bidding operation frequency, bidding time interval, mouse sliding trajectory, keyboard input mode and page dwell time.

[0028] Specifically, the system collects data on suppliers' initial behavior in the bidding auction system in real time, including bid frequency, bid intervals, mouse movement patterns, keyboard input patterns, and page dwell time. This data can also include supplier device information, such as MAC address, IP address, and device model, to further verify the supplier's identity.

[0029] The following is a detailed description of each behavioral data.

[0030] 1. Quotation Operation Frequency: Quotation operation frequency refers to the number of times a supplier submits a quotation per unit time during the bidding process. It reflects the supplier's activeness in participating in the bidding.

[0031] Calculation method: Assume that in a bidding process, the supplier t 1, t 2] Submitted N The frequency of quotation operation is F It can be calculated by the following formula: in, Nis the number of quotes, t 2− t 1 is the total duration of the auction.

[0032] 2. Quotation Interval: The quotation interval refers to the time difference between two consecutive quotations from a supplier. It reflects the supplier's pace of adjusting quotations and decision-making speed.

[0033] Calculation method: Assume that the supplier is at time points t1, t2, t3, …, t N If N quotations are submitted, the time interval Δt between the i-th quotation and the i+1-th quotation is i for: Where i=1,2,…,N−1.

[0034] 3. Mouse Tracks: Mouse tracks are the paths left by users when they move their mouse on a bidding page. They reflect the order in which users browse pages, their focus, and their operating habits.

[0035] Collection Method: The mouse position coordinates (x, y) and timestamp are recorded in real time through front-end code (such as JavaScript) to form a series of trajectory points.

[0036] Analytical significance: ① Hotspot area analysis: By analyzing the dense areas of mouse tracks, you can identify the parts that users are most concerned about (such as price input boxes, competitor quotation information, etc.).

[0037] ②Operation habit analysis: The smoothness and speed of the trajectory can reflect the user's operation proficiency. For example, frequent mouse jitter may indicate that the user is unskilled or nervous.

[0038] ③ Abnormal behavior detection: If the mouse trajectory shows abnormal patterns (such as fast, repetitive straight-line movements), it may indicate that the user is using automated tools.

[0039] 4. Mouse click events: Mouse click events refer to records of mouse clicks performed by users on the bidding page, including the location, time, and frequency of the clicks.

[0040] Collection Method: The front-end code records the coordinates, timestamp, and target element (such as buttons, input boxes, etc.) of each mouse click.

[0041] Analytical significance: ① Operation efficiency analysis: Click frequency and click location can help analyze the efficiency of user operations. For example, frequently clicking the same button may indicate that the user is hesitant or trying different operations.

[0042] ② Focus Analysis: The elements clicked can reflect the features or information that users are most interested in. For example, frequently clicking on competitor quotes may indicate that users are very concerned about the competitor's strategy.

[0043] ③ Abnormal behavior detection: Abnormal click patterns (such as a large number of clicks in a short period of time) may indicate that the user is using scripts or automated tools.

[0044] 5. Keyboard Input Mode: Keyboard input mode refers to the content entered by the user through the keyboard during the bidding process, including bid amount, remarks, etc. At the same time, the input speed, error rate, etc. can also be recorded.

[0045] Collection Method: The front-end code records the characters entered by the keyboard, timestamps, and focus status of the input box.

[0046] Analytical significance: ① Input speed analysis: Input speed can reflect the user's proficiency in operation. Fast input may indicate that the user is very familiar with the system, while slow input may indicate that the user needs more time to think or is not proficient in the operation.

[0047] ② Correlation Analysis: The input content can reflect the correlation between the user input and the actual quote content. A high correlation indicates that the user is less likely to use automated tools, and vice versa.

[0048] ③ Abnormal behavior detection: If the input pattern shows abnormalities (such as no input content or fast input speed but zero correlation), it may indicate that the user is using automated tools.

[0049] 6. Page dwell time: Page dwell time refers to the time a user stays on the bidding page, the total time from entering the page to leaving the page.

[0050] Collection Method: The front-end code records the timestamps of users entering and leaving the page and calculates the length of stay.

[0051] Analytical significance: ① Decision Time Analysis: Dwell time can reflect the time it takes for users to make a quote decision. Shorter dwell times may indicate a quick decision, while longer dwell times may indicate a careful evaluation.

[0052] ② Attention level analysis: A longer stay time may indicate that the user is very concerned about the bidding process, while a shorter stay time may indicate that the user does not care much about the bidding results.

[0053] ③ Abnormal behavior detection: If the dwell time is abnormally short (such as a few seconds), it may indicate that the user is using automated tools to quickly complete the operation without actually browsing the page.

[0054] Step 120: Perform data cleaning and standardization on the initial behavior data, and extract key features to obtain the behavior data.

[0055] Specifically, the initial behavioral data is cleaned and standardized to ensure data quality and consistency. Key features, such as quote time distribution, click heatmaps, and input speed, are extracted from the collected initial behavioral data for analysis.

[0056] Step 200: Input the behavior data into a pre-trained behavior recognition model to obtain a behavior analysis result; wherein, the behavior recognition model is constructed based on historical behavior data through a machine learning algorithm.

[0057] The process of constructing the behavior recognition model in step 200 includes: Step 210: Obtain historical behavior data of multiple suppliers in the bidding auction system, and extract historical key features of the historical behavior data; wherein the historical behavior data includes: normal behavior data and cheating behavior data of each supplier in the multiple suppliers.

[0058] Step 220: Filter and determine target historical key features from the historical key features using a machine learning algorithm; wherein the target historical key features are features that have the greatest impact on behavior recognition.

[0059] Step 230: Build a baseline model based on the target historical key features; wherein the baseline model is used to define a normal behavior pattern.

[0060] Step 240: construct the behavior recognition model based on the baseline model and the target historical key features; wherein the behavior recognition model is used to: analyze the behavior data of the target supplier in real time and identify abnormal behavior patterns.

[0061] In one embodiment, the user's operational data (such as quote frequency, mouse clicks, keyboard input, page dwell time, etc.) is analyzed based on the supplier's historical behavior to determine whether the user has engaged in cheating. The behavior recognition model uses algorithms such as deep learning and random forests to analyze user behavior data, extract features, and build a model.

[0062] 1. Feature extraction: Identify the operating patterns and habits of different suppliers by analyzing historical behavioral data.

[0063] 2. Behavioral pattern recognition: Evaluate the importance of features through algorithms such as random forests and select the features that have the greatest impact on behavior recognition, namely the target's historical key features.

[0064] 3. Abnormal behavior detection: By establishing a baseline model, monitoring the normal behavior patterns of suppliers, identifying abnormal behaviors, and distinguishing between human quotation behaviors and automated tool quotation behaviors, such as frequent quotations and abnormal time intervals, to help discover potential malicious operations.

[0065] Step 300: Determine the risk level of the target supplier based on the behavior analysis results and preset rules, and execute risk control measures corresponding to the risk level.

[0066] It should be noted that the risk levels include: pass, warning and reject.

[0067] Specifically, this embodiment implements risk control logic on the backend, taking appropriate measures based on risk levels. A database is used to record supplier behavior data, risk assessment results, and violation records, facilitating subsequent audits and analysis. Furthermore, risk assessment results and prompts are displayed in real time on the supplier's bidding interface. A graphical verification module is provided for user verification in alert states.

[0068] Specifically, user risk is assessed based on user behavior, categorized as success, caution, and reject. Appropriate intervention measures are then implemented to prevent potential fraud and ensure the fairness and security of the bidding system. Risk assessment includes factors such as bid frequency, bid interval, mouse movement, mouse clicks, keyboard input, and page dwell time.

[0069] Furthermore, combined with the supplier's behavioral data during the bidding process, risk assessment will be conducted from the following dimensions.

[0070] 1. Quotation behavior analysis: Whether the quotation frequency is abnormally high (for example, the number of quotations per unit time is much higher than the normal level).

[0071] Whether the quotation time interval is too short (for example, the quotation is adjusted frequently within a short period of time).

[0072] Whether the quotation trend conforms to normal market rules (such as whether there are abnormal fluctuations).

[0073] 2. Operation behavior analysis: Check whether the mouse sliding trajectory conforms to human operating habits (for example, whether there is straight sliding or abnormal jitter).

[0074] Whether mouse click events are abnormally frequent (for example, whether there are a large number of clicks in a short period of time).

[0075] Whether the keyboard input mode is normal (for example, the input content has low correlation with the actual quotation content, there is no input content during the quotation process, the input speed is fast but the correlation is zero)).

[0076] Whether the page dwell time is too short (for example, whether there is rapid entry and exit behavior).

[0077] 3. Historical behavior analysis: Whether the supplier has been found to have engaged in cheating or unusual behavior in the past.

[0078] Whether the supplier's bidding strategy in historical bidding is consistent with its current behavior.

[0079] Step 300 determines the risk level of the target supplier based on the behavior analysis results and preset rules, and executes risk control measures corresponding to the risk level, including: Step 310: When it is determined that the risk level of the target supplier is a pass, execute a first risk control measure corresponding to the risk level of a pass; wherein the first risk control measure includes: allowing quotation, continuous monitoring, and experience optimization.

[0080] Specifically, when the behavior data of the target supplier fully complies with the normal operating mode and no abnormalities are found, the risk level of the target supplier is determined to be passing.

[0081] The corresponding first risk control measures include: Allow quotation: Suppliers can participate in bidding normally, and the system will not intervene.

[0082] Continuous monitoring: Continue to monitor the supplier's behavioral data in real time to ensure that its subsequent operations remain normal.

[0083] User experience optimization: Based on the normal behavior patterns of suppliers, optimize the bidding page layout and operation process to improve user experience.

[0084] Step 320: When it is determined that the risk level of the target supplier is a warning, a second risk control measure corresponding to the risk level of the warning is executed; wherein the second risk control measure includes: graphic verification.

[0085] Step 320 specifically includes: Step 321: When it is determined that the risk level of the target supplier is a warning, a graphic verification code is sent to the target supplier for verification.

[0086] Step 322: If the target supplier passes the verification within a preset number of times, execute the first risk control measure corresponding to the risk level of pass.

[0087] Step 323: If the target supplier fails verification within a preset number of times, execute the third risk control measure corresponding to the risk level of rejection.

[0088] Specifically, when there are certain anomalies in the target supplier's behavioral data, but not enough to directly determine it as cheating, the target supplier's risk level is determined to be warning.

[0089] Please continue to see Figure 2 , the corresponding second risk control measures include: Graphic verification: Require suppliers to complete a graphic verification code (such as slider verification, click image verification, etc.) to confirm whether they are real users.

[0090] Verification passed: If the user passes the verification, their quotation permissions will be restored and their subsequent behavior will continue to be monitored.

[0091] Verification failure: If a user fails verification multiple times (e.g. 3 times in a row), their risk level will be raised to Reject and corresponding rejection measures will be taken.

[0092] Record warning behaviors: Record warning behaviors in the system log for subsequent analysis and reference.

[0093] Step 330: When it is determined that the risk level of the target supplier is rejection, execute a third risk control measure corresponding to the risk level of rejection; wherein the third risk control measure includes: prohibiting quotation, cheating prompt, recording violation behavior and alarm notification.

[0094] Specifically, when the target supplier's behavior data is obviously abnormal and is determined to be automated tool operation or malicious behavior, the target supplier's risk level is determined to be rejected.

[0095] The corresponding third risk control measures include: Prohibit quotation: Immediately freeze the supplier's quotation rights and terminate its eligibility to participate in the current auction.

[0096] Cheating prompt: A clear cheating prompt message will pop up on the supplier's bidding interface, indicating that their behavior has been judged to be a violation.

[0097] Record violations: Record violations in the system log for subsequent auditing and processing.

[0098] Notify Admin: Sends an alert to the system administrator for further investigation of the violation.

[0099] The above is a description of the steps of the anti-cheating method for a bidding auction system based on behavioral data provided by the present invention. From the description of the above steps, it can be seen that the anti-cheating method for a bidding auction system based on behavioral data provided by the present invention obtains the behavioral data of the target supplier in the bidding auction system; inputs the behavioral data into a pre-trained behavior recognition model to obtain a behavioral analysis result; wherein, the behavior recognition model is: constructed by a machine learning algorithm based on historical behavioral data; determines the risk level of the target supplier according to the behavioral analysis results and preset rules, and executes risk control measures corresponding to the risk level. It can be seen that the present invention monitors the behavioral data of suppliers in the bidding auction system and constructs a behavior recognition model, which can accurately identify whether the supplier's bidding behavior is a human operation, effectively prevent suppliers from automatically quoting through third-party cheating tools, and ensure the fairness and justice of the bidding auction. At the same time, the present invention reduces manual intervention and improves the efficiency and accuracy of anti-cheating through real-time monitoring and automated judgment.

[0100] Based on the above embodiment, in this embodiment, after step 300 determines the risk level of the target supplier based on the behavior analysis results and preset rules and executes risk control measures corresponding to the risk level, the method further includes: Optimizing the behavior recognition model based on the first feedback result and the second feedback result; The first feedback result is obtained by regularly evaluating the risk control measures; and the second feedback result is obtained by analyzing the behavior analysis results to obtain the target cheating method.

[0101] Specifically, we regularly evaluate risk control strategies and adjust judgment rules and thresholds based on actual operating conditions. We also continuously optimize behavioral analysis models based on new cheating methods and technologies to improve the accuracy of risk assessments.

[0102] The anti-cheating method for a bidding auction system based on behavioral data provided in this embodiment improves the accuracy of risk determination by continuously optimizing the behavioral analysis model.

[0103] Based on the above embodiment, in this embodiment, in step 330, if it is determined that the risk level of the target supplier is rejected, after executing the third risk control measure corresponding to the risk level of rejected, the method further includes: Obtaining the appeal result of the target supplier with a risk level of rejected, and re-evaluating based on the appeal result to obtain an evaluation result; The risk control measures are adjusted based on the appeal result and the assessment result.

[0104] Specifically, we provide a user complaint channel, allowing suppliers falsely identified as cheating to submit evidence and request reassessment. Based on user feedback and complaint results, we adjust risk control measures to reduce the false positive rate.

[0105] Furthermore, we regularly audit the implementation of risk control measures to ensure fairness and impartiality. We also conduct statistical analysis of violations and summarize common cheating patterns to provide a basis for subsequent strategy adjustments.

[0106] The anti-cheating method for a bidding auction system based on behavioral data provided in this embodiment reduces the misjudgment rate by adjusting risk control measures.

[0107] The anti-cheating method for auction systems based on behavioral data, provided by an embodiment of the present invention, uses WebSocket or long polling technology to achieve real-time communication between the front-end and back-end, ensuring the timely transmission and processing of behavioral data. It also optimizes the performance of the behavioral analysis model to ensure rapid risk assessment in high-concurrency scenarios. Furthermore, the collected behavioral data is encrypted for transmission and storage to ensure data security and privacy. Users who are mistakenly identified as cheaters are provided with a complaint channel, allowing them to submit evidence and receive a reassessment. The system's behavioral analysis model and decision logic are regularly audited to ensure their accuracy and fairness.

[0108] The anti-cheating method for a bidding auction system based on behavioral data provided by the present invention is described below with reference to specific embodiments.

[0109] Example 1: Normal Quotation Behavior (Success) Background description: Supplier A participates in a bidding activity, and its behavior data is as follows: ① Quotation operation frequency: once every 5 minutes, a total of 3 quotes.

[0110] ② Quotation time interval: Each quotation is about 3 minutes apart, and the time interval is relatively uniform.

[0111] ③ Mouse sliding track: The mouse track is natural, mainly concentrated in the quotation input box, competitor quotation information and submit button area.

[0112] ④ Mouse click event: The click frequency is moderate, mainly clicking the quotation input box and the submit button.

[0113] ⑤Keyboard input mode: The input speed is moderate, the input content includes numerical items related to the quotation amount, there are occasional small modifications, and the input content is highly correlated with the actual quotation content.

[0114] ⑥ Page dwell time: Each dwell time is about 3-5 minutes.

[0115] ⑦IP and MAC address: No change.

[0116] Risk assessment process 1. Quotation behavior analysis: The frequency and time interval of quotations are within the normal range, and there is no frequent quotations or abnormally fast quotations.

[0117] 2. Operation behavior analysis: The mouse trajectory and click behavior are natural, with no abnormal jitter or rapid straight-line movement.

[0118] The keyboard input mode is in line with human operating habits, the input speed is moderate, and the input content is highly correlated with the actual quotation content.

[0119] 3. Comprehensive assessment: Supplier A's behavioral data is completely consistent with normal operating patterns, and no abnormalities were found.

[0120] Risk Level: Result: Success (normal) Action: Allow supplier A to continue bidding, and the system will not intervene.

[0121] Example 2: Suspicious Quotation Behavior (Caution) Background Description Supplier B participates in a bidding activity, and its behavior data is as follows: ① Quotation operation frequency: one quotation every 30 seconds, a total of 12 quotations.

[0122] ② Quotation time interval: The time interval is relatively short, with some intervals being only 30 seconds to 1 minute.

[0123] ③ Mouse sliding trajectory: During the quotation interval, the mouse trajectory is relatively complex, and about 50% of the time it moves frequently in the area outside the quotation page.

[0124] ④ Mouse click event: The click frequency is relatively high, especially concentrated on the quotation input box and submit button.

[0125] ⑤Keyboard input mode: The input speed is faster, but only part of the input contains numeric items, and there are occasional small modifications. The input content is relevant to the actual quotation content.

[0126] ⑥ Page dwell time: Each dwell time is about 2-3 minutes, which is relatively short.

[0127] ⑦IP and MAC address: changed ≥2 times within 1 hour.

[0128] Risk assessment process 1. Quotation behavior analysis: Not to the extent of obvious abnormality.

[0129] There are certain anomalies in the changes of IP and MAC addresses.

[0130] 2. Operation behavior analysis: No obvious abnormalities were found in mouse trajectory and click frequency.

[0131] The keyboard input speed is consistent with human operating habits, and the input content is relevant to the actual quotation content.

[0132] 3. Comprehensive Assessment: Supplier B's behavioral data exhibits certain anomalies, but these anomalies are insufficient to directly determine cheating and require further verification.

[0133] Risk Level Judgment result: Caution measure: Trigger graphic verification and require Supplier B to complete a graphic verification code (such as slider verification).

[0134] If the verification is successful, the quotation authority will be restored; if the verification fails, the risk level will be increased to Reject.

[0135] Example 3: Cheating Behavior (Reject) Background Description Supplier C participates in a bidding activity, and its behavior data is as follows: ① Quotation operation frequency: one quotation per 1 second, a total of 12 quotations.

[0136] ② Quotation time interval: The time interval is extremely short, mostly around 0.5 seconds.

[0137] ③ Mouse sliding track: The mouse track is very small and rarely stays on the quotation page.

[0138] ④ Mouse click events: The click frequency is extremely low, and the mouse rarely stays on the quotation input box and submit button.

[0139] ⑤Keyboard input mode: very little input without any modification, and the input content rarely contains numerical items related to the quotation amount. At the same time, the correlation between the input content and the actual quotation content is zero.

[0140] ⑥ Page dwell time: Each dwell time is only 1 minute, which is obviously too short.

[0141] ⑦IP and MAC address: The number of changes within 1 hour is ≥5 times.

[0142] Risk assessment process 1. Quotation behavior analysis: The quotation frequency is extremely high and the time interval is extremely short, which is obviously inconsistent with normal human operation mode.

[0143] There are obvious anomalies in the changes of IP and MAC addresses.

[0144] 2. Operation behavior analysis: The mouse track is relatively small and rarely stays on the quotation input box and submit button, which is inconsistent with normal operating habits.

[0145] There is very little keyboard input and no modification, the input content rarely contains numerical items related to the quotation amount, and the correlation between the input content and the actual quotation content is zero, which is in line with the operating characteristics of automated tools.

[0146] 3. Comprehensive assessment: Supplier C's behavioral data is clearly abnormal, and we highly suspect it uses automated tools.

[0147] Risk Level Result: Reject measure: Immediately freeze supplier C's quotation authority and block supplier C's quotation in this round.

[0148] A cheating prompt message will pop up on the bidding interface, indicating that the behavior has been judged to be a violation. If cheating behavior is identified again, the system may terminate the participant's eligibility to participate in the current bidding.

[0149] Record violations in the system log for subsequent audit and processing.

[0150] Notify your system administrator for further investigation.

[0151] The anti-cheating method for a bidding auction system based on behavioral data provided by an embodiment of the present invention can accurately identify whether the supplier's bidding behavior is manually operated by monitoring the supplier's behavioral data and system operation data in the bidding auction system, effectively prevent the supplier from automatically bidding through third-party cheating tools, and ensure the fairness and impartiality of the bidding auction; through real-time monitoring and automated judgment, it reduces manual intervention and improves the efficiency and accuracy of anti-cheating; it can effectively reduce cheating behavior in the bidding auction system and improve the security of the system and user experience.

[0152] The following describes the anti-cheating device for a bidding auction system based on behavioral data provided by the present invention. The anti-cheating device for a bidding auction system based on behavioral data described below and the anti-cheating method for a bidding auction system based on behavioral data described above can refer to each other.

[0153] Figure 3 Schematic diagram of the structure of the anti-cheating device for the bidding auction system based on behavioral data provided by the present invention. Figure 3 As shown, the anti-cheating device for a bidding auction system based on behavioral data provided by the present invention includes: The behavior data acquisition module 301 is used to acquire the behavior data of the target supplier in the bidding auction system; The behavior analysis module 302 is used to input the behavior data into a pre-trained behavior recognition model to obtain a behavior analysis result; wherein the behavior recognition model is constructed based on historical behavior data through a machine learning algorithm; The risk identification and control module 303 is used to determine the risk level of the target supplier based on the behavior analysis results and preset rules, and to execute risk control measures corresponding to the risk level.

[0154] The present invention provides an anti-cheating device for a bidding auction system based on behavioral data, which obtains the behavioral data of a target supplier in the bidding auction system; inputs the behavioral data into a pre-trained behavior recognition model to obtain a behavioral analysis result; wherein the behavior recognition model is constructed based on historical behavioral data through a machine learning algorithm; determines the risk level of the target supplier according to the behavioral analysis result and preset rules, and executes risk control measures corresponding to the risk level. It can be seen that the present invention monitors the behavioral data of suppliers in the bidding auction system and constructs a behavior recognition model, which can accurately identify whether the supplier's bidding behavior is a human operation, effectively prevent suppliers from automatically quoting through third-party cheating tools, and ensure the fairness and impartiality of the bidding auction. At the same time, the present invention reduces manual intervention and improves the efficiency and accuracy of anti-cheating through real-time monitoring and automated judgment.

[0155] Based on the above embodiment, in this embodiment, the behavior data acquisition module 301 is specifically used to: Real-time collection of initial behavior data of the target supplier in the bidding auction system; wherein the initial behavior data includes: bidding operation frequency, bidding time interval, mouse sliding trajectory, keyboard input mode and page dwell time; The initial behavior data is cleaned and standardized, and key features are extracted to obtain the behavior data.

[0156] Based on the above embodiment, in this embodiment, the device further includes a building module, specifically configured to: Obtaining historical behavior data of multiple suppliers in a bidding auction system and extracting historical key features of the historical behavior data; wherein the historical behavior data includes: normal behavior data and cheating behavior data of each of the multiple suppliers; Filtering the target historical key features from the historical key features using a machine learning algorithm to determine the target historical key features; wherein the target historical key features are: features that have the greatest impact on behavior recognition; Building a baseline model based on the target's historical key features; wherein the baseline model is used to: define a normal behavior pattern; The behavior recognition model is constructed based on the baseline model and the target historical key features; wherein the behavior recognition model is used to: analyze the behavior data of the target supplier in real time and identify abnormal behavior patterns.

[0157] Based on the above embodiment, in this embodiment, the risk levels include: pass, warning, and reject; The risk identification and control module 303 is specifically used to: If it is determined that the risk level of the target supplier is passing, executing a first risk control measure corresponding to the passing risk level; wherein the first risk control measure includes: allowing quotation, continuous monitoring, and experience optimization; When it is determined that the risk level of the target supplier is a warning, a second risk control measure corresponding to the risk level of the warning is executed; wherein the second risk control measure includes: graphic verification; When it is determined that the risk level of the target supplier is rejection, a third risk control measure corresponding to the risk level of rejection is executed; wherein, the third risk control measure includes: prohibiting quotation, cheating prompt, recording violation behavior and alarm notification.

[0158] Based on the above embodiment, in this embodiment, the risk identification and control module 303 is specifically used to: If the risk level of the target supplier is determined to be a warning, a graphic verification code is sent to the target supplier for verification; If the target supplier passes the verification within a preset number of times, executing the first risk control measure corresponding to the risk level of passing; In the event that the target supplier fails verification within a preset number of times, the third risk control measure corresponding to the risk level of rejection is executed.

[0159] Based on the above embodiment, in this embodiment, the device further includes an optimization module, which is specifically configured to: After determining the risk level of the target supplier according to the behavior analysis results and preset rules and executing risk control measures corresponding to the risk level, the behavior recognition model is optimized based on the first feedback result and the second feedback result; The first feedback result is obtained by regularly evaluating the risk control measures; and the second feedback result is obtained by analyzing the behavior analysis results to obtain the target cheating method.

[0160] Based on the above embodiment, in this embodiment, the device further includes an adjustment module, which is specifically configured to: In the case where it is determined that the risk level of the target supplier is rejected, after executing the third risk control measure corresponding to the risk level of rejected, obtaining the appeal result of the target supplier with the risk level of rejected, and re-evaluating based on the appeal result to obtain an evaluation result; The risk control measures are adjusted based on the appeal result and the assessment result.

[0161] Figure 4 An example of a physical structure diagram of an electronic device is shown below. Figure 4 As shown, the electronic device may be a robot or other electronic device, and may include: a processor 410, a communications interface 420, a memory 430, and a communication bus 440. The processor 410, the communications interface 420, and the memory 430 communicate with each other via the communication bus 440. The processor 410 may call the logic instructions in the memory 430 to execute the anti-cheating method for the bidding auction system based on behavioral data, including: Obtaining the target supplier's behavior data in the bidding auction system; Inputting the behavior data into a pre-trained behavior recognition model to obtain a behavior analysis result; wherein the behavior recognition model is constructed based on historical behavior data through a machine learning algorithm; The risk level of the target supplier is determined based on the behavior analysis results and preset rules, and risk control measures corresponding to the risk level are executed.

[0162] Furthermore, the logic instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the portion that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product, stored in a storage medium, includes instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a mobile hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0163] In another aspect, the present invention further provides a computer program product, comprising a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the anti-cheating method for a bidding auction system based on behavioral data provided by the above methods, including: Obtaining the target supplier's behavior data in the bidding auction system; Inputting the behavior data into a pre-trained behavior recognition model to obtain a behavior analysis result; wherein the behavior recognition model is constructed based on historical behavior data through a machine learning algorithm; The risk level of the target supplier is determined based on the behavior analysis results and preset rules, and risk control measures corresponding to the risk level are executed.

[0164] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon. When executed by a processor, the computer program is implemented to perform the anti-cheating method for a bidding auction system based on behavioral data provided by the above methods, including: Obtaining the target supplier's behavior data in the bidding auction system; Inputting the behavior data into a pre-trained behavior recognition model to obtain a behavior analysis result; wherein the behavior recognition model is constructed based on historical behavior data through a machine learning algorithm; The risk level of the target supplier is determined based on the behavior analysis results and preset rules, and risk control measures corresponding to the risk level are executed.

[0165] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0166] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0167] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A method for preventing cheating in a bidding auction system based on behavioral data, characterized in that: include: Obtaining the target supplier's behavior data in the bidding auction system; Inputting the behavior data into a pre-trained behavior recognition model to obtain a behavior analysis result; wherein the behavior recognition model is constructed based on historical behavior data through a machine learning algorithm; The risk level of the target supplier is determined based on the behavior analysis results and preset rules, and risk control measures corresponding to the risk level are executed.

2. The anti-cheating method for a bidding auction system based on behavioral data according to claim 1 is characterized in that: The obtaining of the target supplier's behavior data in the bidding auction system includes: Real-time collection of initial behavior data of the target supplier in the bidding auction system; wherein the initial behavior data includes: bidding operation frequency, bidding time interval, mouse sliding trajectory, keyboard input mode and page dwell time; The initial behavior data is cleaned and standardized, and key features are extracted to obtain the behavior data.

3. The anti-cheating method for a bidding auction system based on behavioral data according to claim 1, characterized in that: The process of constructing the behavior recognition model includes: Obtaining historical behavior data of multiple suppliers in a bidding auction system and extracting historical key features of the historical behavior data; wherein the historical behavior data includes: normal behavior data and cheating behavior data of each of the multiple suppliers; Filtering the target historical key features from the historical key features using a machine learning algorithm to determine the target historical key features; wherein the target historical key features are: features that have the greatest impact on behavior recognition; Building a baseline model based on the target's historical key features; wherein the baseline model is used to: define a normal behavior pattern; The behavior recognition model is constructed based on the baseline model and the target historical key features; wherein the behavior recognition model is used to: analyze the behavior data of the target supplier in real time and identify abnormal behavior patterns.

4. The anti-cheating method for a bidding auction system based on behavioral data according to claim 1, characterized in that: The risk levels include: pass, warning and reject; Determining the risk level of the target supplier based on the behavior analysis results and preset rules, and executing risk control measures corresponding to the risk level, includes: If it is determined that the risk level of the target supplier is passing, executing a first risk control measure corresponding to the passing risk level; wherein the first risk control measure includes: allowing quotation, continuous monitoring, and experience optimization; When it is determined that the risk level of the target supplier is a warning, a second risk control measure corresponding to the risk level of the warning is executed; wherein the second risk control measure includes: graphic verification; When it is determined that the risk level of the target supplier is rejection, a third risk control measure corresponding to the risk level of rejection is executed; wherein, the third risk control measure includes: prohibiting quotation, cheating prompt, recording violation behavior and alarm notification.

5. The anti-cheating method for a bidding auction system based on behavioral data according to claim 4 is characterized in that: When determining that the risk level of the target supplier is a warning, executing a second risk control measure corresponding to the risk level of the warning includes: If the risk level of the target supplier is determined to be a warning, a graphic verification code is sent to the target supplier for verification; If the target supplier passes the verification within a preset number of times, executing the first risk control measure corresponding to the risk level of passing; In the event that the target supplier fails verification within a preset number of times, the third risk control measure corresponding to the risk level of rejection is executed.

6. The anti-cheating method for a bidding auction system based on behavioral data according to claim 1, characterized in that: After determining the risk level of the target supplier based on the behavior analysis results and preset rules, and executing risk control measures corresponding to the risk level, the method further includes: Optimizing the behavior recognition model based on the first feedback result and the second feedback result; The first feedback result is obtained by regularly evaluating the risk control measures; and the second feedback result is obtained by analyzing the behavior analysis results to obtain the target cheating method.

7. The anti-cheating method for a bidding auction system based on behavioral data according to claim 4 is characterized in that: When determining that the risk level of the target supplier is rejected, after executing the third risk control measure corresponding to the risk level of rejected, the method further includes: Obtaining the appeal result of the target supplier with a risk level of rejected, and re-evaluating based on the appeal result to obtain an evaluation result; The risk control measures are adjusted based on the appeal result and the assessment result.

8. An anti-cheating device for a bidding auction system based on behavioral data, characterized in that: include: A behavior data acquisition module is used to obtain the behavior data of the target supplier in the bidding auction system; A behavior analysis module, configured to input the behavior data into a pre-trained behavior recognition model to obtain behavior analysis results; wherein the behavior recognition model is constructed using a machine learning algorithm based on historical behavior data; The risk identification and control module is used to determine the risk level of the target supplier based on the behavior analysis results and preset rules, and to implement risk control measures corresponding to the risk level.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the anti-cheating method for a bidding auction system based on behavioral data as described in any one of claims 1 to 7 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the anti-cheating method for a bidding auction system based on behavioral data as described in any one of claims 1 to 7 is implemented.

Images