Examination anti-cheating method, device and equipment and storage medium

Abstract

The invention provides an examination anti-cheating method, apparatus and device, and a storage medium. The method comprises the steps of obtaining client use information, answer efficiency information and behavior feature information of an examinee during an online examination; based on the client use information, the answering efficiency information and the behavior feature information, generating comprehensive examination features of the examinee; and judging whether an abnormal behavior exists based on the comprehensive examination characteristics. According to the method and the device, whether the examinee has the abnormal cheating behavior or not is judged by comprehensively considering a plurality of dimension characteristics such as the use condition of the client, the answering efficiency and the behavior characteristics of the examinee in the examination process, so that the cheating behavior in the online examination can be effectively identified, and the fairness of the examination is ensured.

Description

Technical Field

[0001] The present disclosure relates to the technical field of software development, and particularly to an anti-cheating method, device, equipment and storage medium for examinations. Background Art

[0002] With the popularization of online education, online examination systems have been gradually widely applied. However, there are many challenges in online examinations, and anti-cheating is one of the most important issues. At present, most online examination systems adopt the method of video monitoring to prevent cheating, but there is still a problem that cheating behaviors are difficult to completely eliminate. Therefore, how to effectively identify cheating behaviors in online examinations is a technical problem that needs to be solved urgently by those skilled in the art. Summary of the Invention

[0003] In view of this, the present disclosure provides an anti-cheating method, device, equipment and storage medium for examinations, which can effectively identify abnormal behaviors in online examinations to ensure the fairness of examinations.

[0004] According to the first aspect of the present disclosure, there is provided an anti-cheating method for examinations for monitoring examination room behaviors, including:

[0005] Obtaining client usage information, answering efficiency information and behavior characteristic information of a candidate during an online examination;

[0006] Generating a comprehensive examination characteristic of the candidate based on the client usage information, the answering efficiency information and the behavior characteristic information;

[0007] Judging whether there are abnormal behaviors based on the comprehensive examination characteristic.

[0008] In a possible implementation manner, the client usage information includes at least one of screen usage information, network traffic usage information and virtual software usage information.

[0009] In a possible implementation manner, when generating the comprehensive examination characteristic of the candidate based on the client usage information, the answering efficiency information and the behavior characteristic information, it includes:

[0010] Generating a client usage feature vector based on the client usage information;

[0011] Generating an answering efficiency feature vector based on the answering efficiency information;

[0012] Generating a behavior characteristic vector based on the behavior characteristic information;

[0013] Fusing the client usage feature vector, the answering efficiency feature vector and the behavior characteristic vector to generate the comprehensive examination characteristic of the candidate.

[0014] In a possible implementation, before fusing the client usage feature vector, the answering efficiency feature vector, and the behavior feature vector, it further includes an operation of preprocessing the client usage feature vector, the answering efficiency feature vector, and the behavior feature vector.

[0015] In a possible implementation, when determining whether there is any abnormal behavior based on the comprehensive examination features, it includes:

[0016] Obtain the preset benchmark examination features;

[0017] Calculate the similarity between the comprehensive examination features and the benchmark examination features;

[0018] Based on the similarity, determine whether there is any abnormal behavior.

[0019] In a possible implementation, the benchmark examination features are determined based on the comprehensive examination features of non-cheating candidates in historical examination data.

[0020] In a possible implementation, when determining whether there is any abnormal behavior based on the comprehensive examination features, it is obtained based on a pre-trained cheating recognition model.

[0021] According to a second aspect of the present disclosure, there is provided an examination anti-cheating device for monitoring examination room behaviors, including:

[0022] An examination information acquisition module for acquiring client usage information, answering efficiency information, and behavior feature information of a candidate during an online examination;

[0023] An examination feature generation module for generating comprehensive examination features of the candidate based on the client usage information, the answering efficiency information, and the behavior feature information;

[0024] An abnormal behavior recognition module for determining whether there is any abnormal behavior based on the comprehensive examination features.

[0025] According to a third aspect of the present disclosure, there is provided an examination anti-cheating device, including: a processor; a memory for storing processor-executable instructions; wherein, the processor is configured to execute the method described in the first aspect of the present disclosure.

[0026] According to a fourth aspect of the present disclosure, there is provided a non-volatile computer-readable storage medium, on which computer program instructions are stored, wherein, when the computer program instructions are executed by a processor, the method described in the first aspect of the present disclosure is implemented.

[0027] The present disclosure provides an exam anti-cheating method, apparatus, device, and storage medium. The method includes: obtaining client usage information, answering efficiency information, and behavioral characteristic information of a candidate during an online exam; generating a comprehensive exam characteristic of the candidate based on the client usage information, answering efficiency information, and behavioral characteristic information; and determining whether there is any abnormal behavior based on the comprehensive exam characteristic. In the present disclosure, multiple characteristics such as the usage situation of the client, answering efficiency, and behavioral characteristics of the candidate during the exam are comprehensively considered to determine whether the candidate has any abnormal cheating behavior, so as to effectively identify cheating behavior in an online exam and ensure the fairness of the exam.

[0028] Other features and aspects of the present disclosure will become apparent from the following detailed description of exemplary embodiments with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] The drawings included in and constituting a part of this specification, together with the specification, illustrate exemplary embodiments, features, and aspects of the present disclosure and are used to explain the principles of the present disclosure.

[0030] Figure 1 A flowchart showing the exam anti-cheating method according to an embodiment of the present disclosure;

[0031] Figure 2 A schematic block diagram showing the exam anti-cheating apparatus according to an embodiment of the present disclosure;

[0032] Figure 3 A schematic block diagram showing the exam anti-cheating device according to an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0033] Various exemplary embodiments, features, and aspects of the present disclosure will be described in detail below with reference to the accompanying drawings. The same reference numerals in the drawings denote elements having the same or similar functions. Although various aspects of the embodiments are shown in the drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0034] The term "exemplary" used herein means "serving as an example, embodiment, or illustration". Any embodiment described herein as "exemplary" is not necessarily to be construed as superior or better than other embodiments.

[0035] In addition, for a better description of the present disclosure, numerous specific details are given in the following detailed description. Those skilled in the art should understand that the present disclosure can be implemented without some of these specific details. In some instances, methods, means, elements, and circuits well known to those skilled in the art are not described in detail so as to highlight the gist of the present disclosure.

[0036] <Method Embodiment>

[0037] Figure 1 A flowchart showing an exam anti-cheating method according to an embodiment of the present disclosure. As Figure 1 shown, the method includes steps S1100 - S1300.

[0038] S1100, obtain the client usage information, answering efficiency information, and behavioral characteristic information of the examinee during the online exam.

[0039] In a possible implementation, the client usage information includes at least one of screen usage information, network traffic usage information, and virtual software usage information. Among them, the screen usage information includes at least one of display window switching information, application software call information, and web page access information. The network traffic usage information includes at least one of the link information between the exam system and the external server and the usage information of VPN (or proxy). The virtual software usage information includes at least one of the running information of virtual machines and the running information of remote control software.

[0040] In a possible implementation, the answering efficiency information includes at least one of answering time information, question switching information, and question answering order information.

[0041] In a possible implementation, the behavioral characteristic information includes at least one of video monitoring information and voice monitoring information.

[0042] S1200, generate a comprehensive exam characteristic of the examinee based on the client usage information, answering efficiency information, and behavioral characteristic information. The specific steps are as follows:

[0043] First, generate a client usage feature vector based on the client usage information.

[0044] When the client usage information includes screen usage information, a screen usage feature vector will be generated according to the screen usage information, and the screen usage feature vector will be used as a kind of client usage feature vector. Specifically, 1) Obtain the display window switching information in the screen usage information; count the number of display window switches n within the specified time period T according to the switching time of each display window recorded in the display window switching information 1i ; according to the number of window switches n within the specified time period T 1i , calculate the display window switching frequency f within the specified time period T 1i . Among them, f 1i = n 1i / T. 2) Obtain the application software call information in the screen usage information; identify the non-exam application software call information from the application software call information; count the number of calls c of non-exam application software within the specified time period T according to the non-exam application software call information 1i3) Obtain the web access information in the screen usage information; identify the access information of non-examination web pages from the web access information; according to the access information of non-examination web pages, count the total number of accesses w of non-examination web pages within the specified time period T i and the total access duration t 1i 4) For the display window switching frequency f within the specified time period T 1i and the number of calls c of non-examination application software 1i as well as the total number of accesses w of non-examination web pages i and the total access duration t 1i Combine them in order to obtain the screen usage feature vector F screeni Among them, F screeni =(f 1i , c 1i , w i , t 1i ).

[0045] When the client usage information includes network traffic usage information, a network traffic usage feature vector will be generated based on the network traffic usage information, and the network traffic usage feature vector will be used as a type of client usage feature vector. Specifically, 1) Obtain the link information between the examination system and the external server in the network traffic usage information; according to the obtained link information, count the total number of links i between the examination system and the external server within the specified time period T i and the total link traffic b i . 2) Obtain the usage information of VPN (or proxy) in the network traffic usage information, where the usage information of VPN (or proxy) is characterized by the user IP change information and the network connection / disconnection time; according to the user IP change information and the network connection / disconnection time recorded in the usage information of VPN (or proxy), count the total number of user IP changes c within the specified time period T 2i and the total network disconnection duration t 2i . 3) Combine the total number of links i between the examination system and the external server i , the total link traffic b between the examination system and the external server i , the total number of user IP changes c 2i and the total network disconnection duration t 2i in order to obtain the network traffic usage feature vector F networki Among them, F networki =(i i , b i , c 2i , t 2i ).

[0046] When the client usage information includes virtual software usage information, a virtual software usage feature vector will be generated based on the virtual software usage information, and the virtual software usage feature vector will be used as a type of client usage feature vector. Specifically, 1) Obtain the running information of the virtual machine from the virtual software usage information, and count the number of runs and running time s of the virtual machine within the specified time period T according to the running information of the virtual machine 1i . 2) Obtain the running information of the remote software from the virtual software usage information, and count the number of runs and running time t of the remote software within the specified time period T according to the running information of the remote software 3i . 3) Calculate the sum of the number of runs of the virtual machine and the number of runs of the remote software within the specified time period T, and use the calculated sum of the number of runs as the number of uses c of the virtual software within the specified time period T 3i . 4) Combine the number of uses c of the virtual software within the specified time period T 3i , the running time s of the virtual machine 1i and the running time t of the remote software 3i in order to obtain the virtual software usage feature vector F VSi . Among them, F VSi =(c 3i , s 1i , t 3i ).

[0047] It should be noted here that when the client usage information includes screen usage information, network traffic usage information and virtual software usage information at the same time, the generated client usage feature vector will include the screen usage feature vector F screeni , the network traffic usage feature vector F networki and the virtual software usage feature vector F VSi .

[0048] Second, generate a test answering efficiency feature vector based on the test answering efficiency information. Specifically, 1) Extract the test answering time information from the test answering efficiency information. The test answering time information includes the start test answering time, the test submission time and the actual test answering time for each question. Among them, the actual test answering time for each question is equal to the time difference between switching to the current question and clicking to switch to the next question. Calculate the time difference between the start test answering time and the test submission time, and divide it by the total number of questions in the test paper to obtain the average test answering time t of the candidate 4i . Calculate the cumulative value of the actual test answering time for each question answered within the specified time period T to obtain the actual test answering time t of the candidate within the specified time period T 5i . Calculate the ratio t 5i of the actual test answering time t 4i to the average test answering time t 5i / t 4i2) Extract the question switching information from the answering efficiency information. The question switching information includes the number of question switches within a specified time period T. Calculate the question replacement frequency f within the specified time period T based on the number of question switches within the specified time period T. 2i , the question switching frequency f 2i = number of question switches / specified time period T. 3) Extract the question answering order information from the answering efficiency information. The question answering order information records the answering order when the candidate answers questions normally, that is, the answering serial number increases by 1 each time, and also records the answering order during the actual exam. Calculate the sum of the answering order serial numbers when the candidate answers questions normally within the specified time period T as the first question number accumulation value. Calculate the sum of the answering order serial numbers when the candidate actually answers questions within the specified time period T as the second question number accumulation value. Calculate the difference s 2i . Calculate the difference s 2i and the ratio of the specified time period T to obtain the question browsing mode feature within the specified time period T, that is, the question browsing mode feature within the specified time period T = s 2i / T. 4) Combine the ratio t 5i of the actual answering time t within the specified time period T to the average answering time t 4i , the question replacement frequency f 5i / t 4i , and the question browsing mode feature = s 2i / T in sequence to obtain the answering efficiency feature vector F 2i within the specified time period T. Among them, F speedi = (t speedi / t 5i , f 4i , s 2i / T). 2i

[0049] Third, generate a behavior feature vector based on the behavior feature information. Specifically, 1) Extract the video surveillance information from the behavior feature information, and call the AI recognition interface for abnormal image analysis to analyze the behavior features of the candidate's facial expression, eye direction, body sitting posture, etc. If the candidate has abnormal behaviors, such as frequent turning of the head, looking at places other than the screen, etc., identify the picture with abnormal behaviors as an abnormal picture, and count the number of abnormal pictures m i that appear within the specified time period T. 2) Extract the voice surveillance information from the behavior feature information, and call the AI recognition interface for abnormal sound analysis to analyze the environmental noise around the candidate. If there are voices of other people or other abnormal noises in the environment around the candidate, identify it as an abnormal sound, and count the number of abnormal sounds n 2i that appear within the specified time period T. 3) The number of abnormal pictures m within the specified time periodi and abnormal sound n 2i Combine them in order to obtain the behavior feature vector F within the specified time period T actioni . Among them, F actioni =(m i , n 2i ).

[0050] Fourth, fuse the client usage feature vector, the answering efficiency feature vector, and the behavior feature vector to generate the comprehensive examination features of the candidate. Specifically, sort the client usage feature vector, the answering efficiency feature vector, and the behavior feature vector within the specified time period T in order, and the comprehensive examination features of the candidate within the specified time period T can be generated. For example, the client usage feature vector within the specified time period T will include the screen usage feature vector F screeni , the network traffic usage feature vector F networki , and the virtual software usage feature vector F VSi . When the answering efficiency feature vector is F speed , and the behavior feature vector is F action , the comprehensive examination features F exami of the candidate within the specified time period T=(F screeni , F networki , F VSi , F speedi , F actioni ).

[0051] It should be noted here that the specified time period T can be the entire examination period or multiple monitoring periods of equal duration divided according to a preset sampling time interval, and no specific limitation is made here. When the specified time period T is multiple monitoring periods of equal duration divided according to a preset sampling time interval, a corresponding comprehensive examination feature F exami of the candidate will be obtained in each detection period in the above manner, that is, the number of monitoring periods into which the entire examination process is divided is the number of comprehensive examination features F exami of the candidate that will be generated in sequence. The multiple comprehensive examination features F exami of the candidate are arranged in chronological order, and the comprehensive examination features F exam of the candidate in the entire examination process are obtained, that is, F exam =(F exam1 , F exam2 , ……, F examD ), where D is the total number of comprehensive examination features of the candidate calculated during the entire examination process.

[0052] In a possible implementation, in order to standardize different types of data so that they have the same scale and facilitate subsequent fusion analysis, before fusing the client usage feature vector, the answering efficiency feature vector, and the behavior feature vector, it also includes a preprocessing operation of standardizing the client usage feature vector, the answering efficiency feature vector, and the behavior feature vector. Specifically, traverse each feature vector. For the currently traversed feature vector, count the maximum and minimum values of the current feature vector, and then perform Min-Max standardization processing on the current feature vector based on the maximum and minimum values of the current feature vector to obtain the standardized feature vector of the current feature vector. After the traversal ends, the standardized feature vectors of each feature vector can be obtained. After obtaining the standardized feature vectors of each feature, fuse them to generate the corresponding comprehensive examination features of the candidates.

[0053] In a possible implementation, before the preprocessing operation of standardizing the client usage feature vector, the answering efficiency feature vector, and the behavior feature vector, it also includes a data cleaning operation. Through data cleaning, abnormal data generated due to network fluctuations can be eliminated. After the complete data cleaning, the preprocessing operation of standardization is performed, thereby improving the accuracy of the calculation of the comprehensive examination features of the candidates.

[0054] S1300, based on the comprehensive examination features, determine whether there is abnormal behavior. Among them, the abnormal behavior here includes the abnormal behavior of exam cheating. Specifically, it may include the following steps:

[0055] First, obtain the reference examination features. Specifically, obtain the comprehensive examination features of all candidates taking the same online exam within the specified time period T and calculate the average value of the comprehensive examination features of all candidates. Take the calculated average value as the reference examination feature F within the specified time period T baselinei . Among them, the reference examination feature F baselinei The calculation formula is as follows:

[0056]

[0057] In the formula, N is the number of all candidates taking the same online exam, is the comprehensive examination feature of the kth candidate within the specified time period T.

[0058] When the specified time period T is divided into multiple monitoring periods with equal durations according to the preset sampling time interval, a corresponding reference examination feature F will be obtained in each detection period in the above manner baselinei , that is, the entire exam process is divided into as many monitoring periods as there are, and as many reference examination features F will be generated in sequence baselinei , multiple reference examination features F baselineiArrange them in chronological order, and the benchmark examination characteristics F of the examinee during the entire examination process can be obtained. baseline , that is, F baseline =(F baseline1 , F baseline2 , ……, F baselineD ), where D is the total number of the calculated benchmark examination characteristics F baselinei during the entire examination process.

[0059] Second, calculate the similarity between the comprehensive examination characteristics and the benchmark examination characteristics. Specifically, the calculation formula for the similarity is as follows:

[0060]

[0061] Third, based on the similarity, determine whether there is cheating behavior in the examination. Specifically, determine whether the similarity is greater than a preset similarity threshold. When it is less than the similarity threshold, it is determined that the examinee has cheating behavior; otherwise, it is considered that the examinee does not have cheating behavior.

[0062] In another possible implementation manner, when determining whether an examinee has cheating behavior based on the comprehensive examination characteristics, it is obtained based on a pre-trained cheating recognition model. Specifically, obtain a preset number of comprehensive examination characteristics, and manually label whether the corresponding label of each comprehensive examination characteristic is cheating or not cheating. Train a pre-loaded binary classification model based on each comprehensive examination characteristic and the corresponding label, so as to obtain a cheating recognition model that can determine whether an examinee cheats based on the comprehensive examination characteristics. After obtaining the cheating recognition model, input the comprehensive examination characteristics of the examinee into the cheating recognition model, and then it can be determined whether there is cheating behavior in the examination through this cheating recognition model.

[0063] The present disclosure provides an examination anti-cheating method, including: obtaining the client usage information, answering efficiency information, and behavior characteristic information of an examinee during an online examination; generating the comprehensive examination characteristics of the examinee based on the client usage information, answering efficiency information, and behavior characteristic information; and determining whether there is abnormal behavior based on the comprehensive examination characteristics. In the present disclosure, when determining whether an examinee has abnormal cheating behavior, multiple characteristics such as the usage situation of the client, answering efficiency, and behavior characteristics of the examinee during the examination process are comprehensively considered, so as to effectively identify cheating behavior in an online examination and ensure the fairness of the examination.

[0064] To clearly illustrate the technical solution of the present disclosure, the technical solution of the present disclosure will be described again below with a specific example. Specifically, the examination anti-cheating method includes the following steps:

[0065] First, multi-factor identity authentication is performed on the examinees. Specifically, when an examinee enters the cloud examination system, they need to first pass the first legality verification through the pre-assigned admission ticket number and password. After the first legality verification is passed, the examination system is entered. Then, the examinee undergoes face recognition through the camera. The examination system compares the real-time captured facial image of the examinee with the facial image at the time of registration to determine whether it is the examinee himself / herself. If it is determined to be the examinee himself / herself, the second legality verification is passed. At this time, the examinee can enter the examination system interface to take the exam; if it is determined not to be the examinee himself / herself, the second legality verification fails, and the examinee is prompted to re-perform face recognition or the invigilator is notified for manual verification. During the examination process, random legality verification is also carried out by randomly capturing the facial image of the examinee and comparing it with the facial image at the time of registration. If it is found that the captured facial image of the examinee is inconsistent with the facial image at the time of registration, the random legality verification fails, and the invigilator is notified for manual verification.

[0066] Second, after the examinee enters the examination system, the examination system automatically activates the monitoring mode, activates monitoring devices such as the camera, microphone, screen monitoring software, and network traffic monitoring system, and starts collecting the client usage information, answering efficiency information, and behavioral characteristic information of the examinee according to a preset time period.

[0067] Third, for each time period, based on the collected client usage information, answering efficiency information, and behavioral characteristic information, a comprehensive examination feature F of the examinee is generated. exami After the examination ends, the various comprehensive examination features F calculated during the examination exami are sorted in order to obtain the comprehensive examination feature F of the examinee during the entire examination process. exam

[0068] Fourth, calculate the similarity between the comprehensive examination feature F exam and the reference examination feature F baseline and determine whether there is any abnormal cheating behavior of the examinee during the examination based on the calculated similarity.

[0069] <Device Embodiment>

[0070] Figure 2 The schematic block diagram of an examination anti-cheating device according to an embodiment of the present disclosure is shown.

[0071] As Figure 2 shown, the device 100 includes:

[0072] An examination information acquisition module 110, configured to acquire the client usage information, answering efficiency information, and behavioral characteristic information of the examinee during the online examination;

[0073] An exam feature generation module 120, configured to generate comprehensive exam features of examinees based on client usage information, answering efficiency information, and behavioral feature information;

[0074] An abnormal behavior recognition module 130, configured to determine whether there is any abnormal behavior based on the comprehensive exam features.

[0075] <Device Embodiment>

[0076] Figure 3 A schematic block diagram of an exam anti-cheating device according to an embodiment of the present disclosure is shown. As Figure 3 shown, the exam anti-cheating device 200 includes: a processor 210 and a memory 220 for storing executable instructions that can be executed by the processor 210. Among them, the processor 210 is configured to implement the exam anti-cheating method described in any one of the foregoing when executing the executable instructions.

[0077] Here, it should be noted that the number of processors 210 can be one or more. At the same time, in the exam anti-cheating device 200 of the embodiment of the present disclosure, an input device 230 and an output device 240 may further be included. Among them, the processor 210, the memory 220, the input device 230, and the output device 240 may be connected through a bus or in other ways, which is not specifically limited herein.

[0078] The memory 220, as a computer-readable storage medium, can be used to store software programs, computer-executable programs, and various modules, such as: programs or modules corresponding to the exam anti-cheating method of the embodiment of the present disclosure. The processor 210 executes various functional applications and data processing of the exam anti-cheating device 200 by running the software programs or modules stored in the memory 220.

[0079] The input device 230 can be used to receive input numbers or signals. Among them, the signal can be a key signal related to user settings and function control of the device / terminal / server. The output device 240 may include a display device such as a display screen.

[0080] <Storage Medium Embodiment>

[0081] According to a fourth aspect of the present disclosure, a non-volatile computer-readable storage medium is further provided, on which computer program instructions are stored, and when the computer program instructions are executed by the processor 210, the exam anti-cheating method described in any one of the foregoing is implemented.

[0082] The embodiments of the present disclosure have been described above. The above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations are obvious to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The choice of terms used herein is intended to best explain the principles of the embodiments, the practical application, or the technical improvements in the technology in the market, or to enable other ordinary skilled persons in the art to understand the embodiments disclosed herein.

Claims

1. A method for preventing cheating in an exam, characterized in that, For monitoring the behavior in the examination room, including: Obtaining the client usage information, answering efficiency information, and behavioral characteristic information of the examinee during the online examination; Generating the comprehensive examination characteristics of the examinee based on the client usage information, the answering efficiency information, and the behavioral characteristic information; Judging whether there is any abnormal behavior based on the comprehensive examination characteristics.

2. The method according to claim 1, wherein The client usage information includes at least one of screen usage information, network traffic usage information, and virtual software usage information.

3. The method according to claim 1, wherein When generating the comprehensive examination characteristics of the examinee based on the client usage information, the answering efficiency information, and the behavioral characteristic information, it includes: Generating a client usage feature vector based on the client usage information; Generating an answering efficiency feature vector based on the answering efficiency information; Generating a behavioral characteristic vector based on the behavioral characteristic information; Fusing the client usage feature vector, the answering efficiency feature vector, and the behavioral characteristic vector to generate the comprehensive examination characteristics of the examinee.

4. The method according to claim 1, wherein Before fusing the client usage feature vector, the answering efficiency feature vector, and the behavioral characteristic vector, it also includes an operation of preprocessing the client usage feature vector, the answering efficiency feature vector, and the behavioral characteristic vector.

5. The method according to claim 1, characterized in that, When judging whether there is any abnormal behavior based on the comprehensive examination characteristics, it includes: Obtaining the preset benchmark examination characteristics; Calculating the similarity between the comprehensive examination characteristics and the benchmark examination characteristics; Judging whether there is any abnormal behavior based on the similarity.

6. The method according to claim 5, characterized in that, The benchmark examination characteristics are determined based on the comprehensive examination characteristics of the examinees without cheating in the historical examination data.

7. The method according to claim 1, wherein When judging whether there is any abnormal behavior based on the comprehensive examination characteristics, it is obtained based on a pre-trained cheating recognition model.

8. An exam anti-cheating device, characterized in that, For monitoring the behavior in the examination room, including: An examination information acquisition module for obtaining the client usage information, answering efficiency information, and behavioral characteristic information of the examinee during the online examination; An examination characteristic generation module for generating the comprehensive examination characteristics of the examinee based on the client usage information, the answering efficiency information, and the behavioral characteristic information; An abnormal behavior recognition module for judging whether there is any abnormal behavior based on the comprehensive examination characteristics.

9. An exam anti-cheating device, characterized in that, Including: A processor; A memory for storing the executable instructions that can be executed by the processor; Wherein, the processor is configured to implement the method described in any one of claims 1 to 7 when executing the executable instructions.

10. A non-volatile computer-readable storage medium having computer program instructions stored thereon, characterized in that, The computer program instructions implement the method described in any one of claims 1 to 7 when executed by the processor.

Images