Risk level prediction method, device, equipment and medium based on micro-expression recognition

Abstract

The present invention relates to the field of artificial intelligence, and provides a risk level prediction method, device, equipment and medium based on micro-expression recognition, which can randomly generate more easy-to-understand questions in real time according to the initial risk level and the extracted questions, effectively avoid malicious guidance of relevant personnel, ensure the accuracy of risk assessment results, automatically adjust the initial risk level according to the credibility of model prediction and the user's answer, make the level determination more accurate, automatically generate a reasonable number of questions according to the total amount of generated questions, avoid wasting time costs, combine micro-expression recognition to determine the credibility of user answers, ensure that the collected user answers are from the user's real feedback, and dynamically adjust the risk level at the same time, so that the prediction of the risk level is more automated, and the prediction results are more accurate. In addition, the present invention also relates to blockchain technology, and the prediction model can be stored in a blockchain node.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a risk level prediction method, device, equipment and medium based on micro-expression recognition. Background Art

[0002] In the financial sector, sales agencies mainly sell various financial products to retail users. According to regulatory requirements, sales agencies need to comply with the principle of investor suitability, that is, they must sell suitable products to suitable investors. Under the guidance of relevant policies, sales agencies have carried out a lot of work around product classification, investor classification, risk assessment, and audio and video recording of investment processes.

[0003] For example, when making a graded prediction of investors' risk tolerance, it is mainly done through offline and online scenarios: in the offline scenario, institutional staff provide a standard paper questionnaire, and the investor checks the box in writing. The institutional staff then scans the answer sheet into the system, and the system automatically completes the scoring and determines the investor's risk level based on the answers to the multiple-choice questions; in the online scenario, the investor enters the risk assessment page in the mobile APP (Application), clicks to complete a set of multiple-choice questions, and the APP server backend system automatically calculates the investor's risk level based on the answers to the multiple-choice questions.

[0004] Both of the above scenarios use static scoring of multiple-choice answers to grade investors’ risk preferences, which has the following main defects:

[0005] (1) The written expressions in the financial industry are relatively obscure and difficult to understand. In order to complete the questionnaire, the respondents need to read long paragraphs of text, which makes the whole process experience poor.

[0006] (2) Some investors can distort the assessment results by artificially selecting specific answers. Some institutional sales personnel have the opportunity to induce investors to select specific answers, thereby making originally unsuitable investors eligible to purchase the products they recommend, resulting in inaccurate final risk assessment results. Summary of the invention

[0007] In view of the above, it is necessary to provide a risk level prediction method, device, equipment and medium based on micro-expression recognition, aiming to solve the problem of inaccurate prediction of the risk level that a user can bear.

[0008] A risk level prediction method based on micro-expression recognition, the risk level prediction method based on micro-expression recognition comprising:

[0009] When it is detected that a user has triggered an answering instruction, the user identity information is obtained according to the answering instruction;

[0010] Randomly configuring an initial risk level of the user;

[0011] Continuously extracting questions from the configured question bank, and generating questions according to the initial risk level and the extracted questions;

[0012] Announce each generated question, and collect the user's micro-expression features and the user's feedback answer during the announcement process;

[0013] Constructing features corresponding to each question according to the initial risk level, the user identity information, the micro-expression features and the answers;

[0014] Calling a pre-trained prediction model, inputting the features corresponding to each question into the prediction model, and obtaining output data of the prediction model as the credibility corresponding to each question;

[0015] Adjusting the initial risk level according to the credibility of each question and the answer to each question;

[0016] Detect the total amount of generated questions in real time, and stop generating questions when the total amount and the credibility of each question reach the configured conditions;

[0017] The risk level of the user is determined according to the credibility corresponding to each question.

[0018] According to a preferred embodiment of the present invention, generating a question based on the initial risk level and the extracted question surface includes:

[0019] Acquiring risk scenarios from the configuration question bank according to the initial risk level;

[0020] Get configuration options;

[0021] The extracted question surface, the risk scenario and the configuration options are combined to obtain the question.

[0022] According to a preferred embodiment of the present invention, collecting the micro-expression features of the user and the answer fed back by the user during the reporting process includes:

[0023] Invoking a micro-expression recognition model to recognize the micro-expression of the user to obtain the micro-expression feature;

[0024] Starting a preset audio and video acquisition device to collect the voice input by the user;

[0025] The speech input by the user is semantically recognized to obtain an answer fed back by the user.

[0026] According to a preferred embodiment of the present invention, before calling the pre-trained prediction model, the method further includes:

[0027] Get historical answer records;

[0028] Constructing a training sample according to the historical answer record, wherein the training sample includes at least one question sample, a feature corresponding to each question sample, and a credibility corresponding to each question sample;

[0029] The prediction model is obtained by performing unsupervised learning using the training samples.

[0030] According to a preferred embodiment of the present invention, adjusting the initial risk level according to the credibility corresponding to each question and the answer corresponding to each question includes:

[0031] Get configuration credibility;

[0032] For each question, when the credibility corresponding to the question is greater than the configured credibility and the answer corresponding to the question is a negative type, reduce the initial risk level according to the configured step size; or

[0033] When the credibility corresponding to the question is greater than the configured credibility, and the answer corresponding to the question is of the affirmative type, the initial risk level is increased according to the configured step size; or

[0034] When the credibility corresponding to the question is greater than the configuration credibility, and the answer corresponding to the question is of an uncertain type, the initial risk level is maintained unchanged; or

[0035] When the credibility corresponding to the question is less than or equal to the configuration credibility, the initial risk level is kept unchanged.

[0036] According to a preferred embodiment of the present invention, the method further comprises:

[0037] Acquire, from the generated topics, the number of topics whose credibility is greater than the configured credibility as a first number;

[0038] When the total amount is greater than or equal to a first threshold, and / or the first number is greater than or equal to a second threshold, it is determined that the total amount and the credibility corresponding to each question meet the configuration condition.

[0039] According to a preferred embodiment of the present invention, determining the risk level of the user according to the credibility corresponding to each question includes:

[0040] Acquire, from the generated topics, topics whose credibility is greater than the configured credibility as candidate topics;

[0041] Obtaining the topic with the highest corresponding risk level from the candidate topics as the target topic;

[0042] The risk level corresponding to the target topic is determined as the risk level of the user.

[0043] A risk level prediction device based on micro-expression recognition, the risk level prediction device based on micro-expression recognition comprising:

[0044] An acquisition unit, configured to acquire user identity information according to the answering instruction when detecting that a user has triggered an answering instruction;

[0045] a configuration unit, configured to randomly configure the initial risk level of the user;

[0046] A generating unit, configured to continuously extract questions from a configured question bank, and generate questions according to the initial risk level and the extracted questions;

[0047] A collection unit, used to announce each generated question and collect the user's micro-expression features and the answer fed back by the user during the announcement process;

[0048] A construction unit, configured to construct a feature corresponding to each question according to the initial risk level, the user identity information, the micro-expression feature and the answer;

[0049] A calling unit, used to call a pre-trained prediction model, input the features corresponding to each question into the prediction model, and obtain output data of the prediction model as the credibility corresponding to each question;

[0050] An adjustment unit, configured to adjust the initial risk level according to the credibility corresponding to each question and the answer corresponding to each question;

[0051] A detection unit, used for detecting the total amount of generated questions in real time, and stopping the generation of questions when the total amount and the credibility corresponding to each question reach a configured condition;

[0052] The determination unit is used to determine the risk level of the user according to the credibility corresponding to each question.

[0053] A computer device, comprising:

[0054] a memory storing at least one instruction; and

[0055] The processor executes the instructions stored in the memory to implement the risk level prediction method based on micro-expression recognition.

[0056] A computer-readable storage medium stores at least one instruction, and the at least one instruction is executed by a processor in a computer device to implement the risk level prediction method based on micro-expression recognition.

[0057] It can be seen from the above technical solutions that the present invention can combine micro-expression recognition to determine the credibility of user answers, ensure that the collected user answers are real feedback from the user, and dynamically adjust the risk level to make the prediction of the risk level more automated and the prediction results more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 It is a flow chart of a preferred embodiment of the risk level prediction method based on micro-expression recognition of the present invention.

[0059] Figure 2 It is a functional module diagram of a preferred embodiment of the risk level prediction device based on micro-expression recognition of the present invention.

[0060] Figure 3 It is a structural schematic diagram of a computer device of a preferred embodiment of the present invention for implementing a risk level prediction method based on micro-expression recognition. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is described in detail below with reference to the accompanying drawings and specific embodiments.

[0062] like Figure 1 FIG. 1 is a flow chart of a preferred embodiment of the risk level prediction method based on micro-expression recognition of the present invention. According to different requirements, the order of the steps in the flow chart can be changed, and some steps can be omitted.

[0063] The risk level prediction method based on micro-expression recognition is applied to one or more computer devices, which are devices that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and their hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0064] The computer device may be any electronic product that can perform human-computer interaction with a user, such as a personal computer, a tablet computer, a smart phone, a personal digital assistant (PDA), a game console, an interactive network television (IPTV), a smart wearable device, etc.

[0065] The computer device may also include a network device and / or a user device, wherein the network device includes, but is not limited to, a single network server, a server group consisting of multiple network servers, or a cloud consisting of a large number of hosts or network servers based on cloud computing.

[0066] The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDN), as well as big data and artificial intelligence platforms.

[0067] Among them, Artificial Intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0068] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0069] The network where the computer device is located includes but is not limited to the Internet, a wide area network, a metropolitan area network, a local area network, a virtual private network (VPN), etc.

[0070] S10, when it is detected that a user has triggered an answering instruction, the user identity information is obtained according to the answering instruction.

[0071] In this embodiment, the users may include investors and the like.

[0072] In this embodiment, when it is detected that the user logs into the configuration page or application, it can be determined that the user has triggered the question answering instruction.

[0073] In this embodiment, the user identity information may include, but is not limited to, a combination of one or more of the following information: permanent location, gender, and age.

[0074] The user identity information is used to represent the identity of the user and is non-sensitive information.

[0075] In this embodiment, obtaining the user identity information according to the answering instruction includes:

[0076] Obtaining login information of the user corresponding to the answer instruction;

[0077] The user identity information is determined according to the login information.

[0078] For example, the user's username can be determined based on the question-answering instruction, and the user's permanent region, gender, age and other information can be further determined based on the username as the user identity information.

[0079] S11, randomly configuring the initial risk level of the user.

[0080] The initial risk level may be configured according to actual mission requirements. For example, the initial risk level may include, but is not limited to, aggressive, growth, balanced, stable, and conservative.

[0081] In this embodiment, the initial risk level may be randomly configured for the user, and then dynamically adjusted.

[0082] S12, continuously extracting questions from the configured question bank, and generating questions according to the initial risk level and the extracted questions.

[0083] The configuration question bank can be maintained in advance and updated in real time according to actual needs to meet the latest test needs.

[0084] Traditional subjective multiple-choice questions include many financial terms, and the textual expressions are relatively obscure and difficult to understand. In order to complete the questionnaire, investors need to read long paragraphs of text, and the whole process experience is poor.

[0085] For example, a traditional topic could be:

[0086] “When your investments fluctuate to what extent, you will become significantly anxious

[0087]

[0088] Since the questions are fixed, some investors can artificially choose specific answers to distort the evaluation results. Some institutional sales personnel also have the opportunity to induce investors to choose specific answers, thereby making originally unsuitable investors eligible to purchase their recommended products, resulting in inaccurate final risk assessment results.

[0089] In order to solve the above problem, this embodiment can automatically generate questions in real time.

[0090] In this embodiment, generating a question based on the initial risk level and the extracted question surface includes:

[0091] Acquiring risk scenarios from the configuration question bank according to the initial risk level;

[0092] Get configuration options;

[0093] The extracted question surface, the risk scenario and the configuration options are combined to obtain the question.

[0094] For example, the question is "You have held a stock fund for a year and suddenly made a large provision for losses, resulting in losses", and the initial risk level is "conservative", and the risk scenario corresponding to "conservative" is "loss of principal within 10%", then after combining the question and risk scenario, the question is:

[0095] "You have held a stock fund for a year and suddenly made a large provision for losses, with a 10% loss. Do you show obvious anxiety about this?

[0096] A. Yes

[0097] B. No

[0098] C. It depends on the situation, not necessarily."

[0099] In the above embodiment, the questions are automatically generated by matching the risk level. The generated questions not only include the title, but also incorporate the risk scenarios related to the risk level, making the questions more understandable. At the same time, since the generated questions are random in real time, the malicious guidance of relevant personnel is effectively avoided, ensuring the accuracy of the risk assessment results.

[0100] S13, reporting each generated question, and collecting the user's micro-expression features and the answer fed back by the user during the reporting process.

[0101] Specifically, each generated question can be converted into voice, and each generated question can be voice broadcasted to facilitate user evaluation.

[0102] In this embodiment, collecting the user's micro-expression features and the user's feedback answer during the broadcast process includes:

[0103] Invoking a micro-expression recognition model to recognize the micro-expression of the user to obtain the micro-expression feature;

[0104] Starting a preset audio and video acquisition device to collect the voice input by the user;

[0105] The speech input by the user is semantically recognized to obtain an answer fed back by the user.

[0106] The micro-expression recognition model may adopt a facial recognition component module commonly used in the market, and the present invention does not limit the micro-expression recognition model.

[0107] The micro-expression features are used to reflect the user's emotions, as shown in the following table:

[0108]

[0109] The preset audio and video acquisition device may include a camera or a microphone of a terminal that broadcasts the question.

[0110] Among them, a general semantic recognition model can be used to perform semantic recognition on the voice input by the user, and the present invention is not limited thereto.

[0111] S14, constructing features corresponding to each question according to the initial risk level, the user identity information, the micro-expression features and the answers.

[0112] For example: the initial risk level can be configured as R (the values ​​are: aggressive R1, growth R2, balance R3, robust R4, conservative R5), the user identity information can be configured as region D, gender G, age O, the micro-expression features can be configured as T1...Tn, the answer can be configured as A (the values ​​can be yes, no, uncertain), and all the above features can be combined into a feature vector to obtain the features corresponding to each question.

[0113] S15, calling a pre-trained prediction model, inputting the features corresponding to each question into the prediction model, and obtaining the output data of the prediction model as the credibility corresponding to each question.

[0114] In at least one embodiment of the present invention, before calling the pre-trained prediction model, the method further includes:

[0115] Get historical answer records;

[0116] Constructing a training sample according to the historical answer record, wherein the training sample includes at least one question sample, a feature corresponding to each question sample, and a credibility corresponding to each question sample;

[0117] The prediction model is obtained by performing unsupervised learning using the training samples.

[0118] The prediction model may be any network model with prediction function, such as a neural network model, etc., which is not limited by the present invention.

[0119] Through the above implementation, a prediction model can be automatically constructed to achieve automatic prediction of credibility, thereby improving the accuracy of the prediction.

[0120] Furthermore, after the prediction model is trained, the features corresponding to each question can be input into the prediction model, and the output data of the prediction model can be obtained as the credibility corresponding to each question. For example, the credibility corresponding to each question can be a percentage, such as 70%.

[0121] S16, adjusting the initial risk level according to the credibility corresponding to each question and the answer corresponding to each question.

[0122] Since the initial risk level is obtained by random initialization, in order to obtain an accurate risk level, the initial risk level needs to be adjusted.

[0123] In this embodiment, adjusting the initial risk level according to the credibility corresponding to each question and the answer corresponding to each question includes:

[0124] Get configuration credibility;

[0125] For each question, when the credibility corresponding to the question is greater than the configured credibility and the answer corresponding to the question is a negative type, reduce the initial risk level according to the configured step size; or

[0126] When the credibility corresponding to the question is greater than the configured credibility, and the answer corresponding to the question is of the affirmative type, the initial risk level is increased according to the configured step size; or

[0127] When the credibility corresponding to the question is greater than the configuration credibility, and the answer corresponding to the question is of an uncertain type, the initial risk level is maintained unchanged; or

[0128] When the credibility corresponding to the question is less than or equal to the configuration credibility, the initial risk level is kept unchanged.

[0129] For example, when the credibility corresponding to the topic is 60% and the configuration credibility is 70%, since the credibility corresponding to the topic is less than the configuration credibility, the initial risk level is kept unchanged.

[0130] In the above embodiment, the initial risk level is automatically adjusted according to the credibility of the model prediction and the user's answer, so that the determination of the risk level is more accurate.

[0131] S17, detecting the total amount of generated questions in real time, and stopping generating questions when the total amount and the credibility corresponding to each question reach the configured conditions.

[0132] It is understandable that when the number of generated questions is too large, it will waste the time of the person being evaluated and increase the time cost of the evaluation; when the number of generated questions is too small, the evaluation results will be inaccurate due to the small number of evaluation questions and insufficient base. Therefore, it is necessary to reasonably determine the number of generated questions.

[0133] Furthermore, the method further comprises:

[0134] Acquire, from the generated topics, the number of topics whose credibility is greater than the configured credibility as a first number;

[0135] When the total amount is greater than or equal to a first threshold, and / or the first number is greater than or equal to a second threshold, it is determined that the total amount and the credibility corresponding to each question meet the configuration condition.

[0136] The first threshold and the second threshold can be custom configured, for example, the first threshold can be 10, and the second threshold can be 5.

[0137] When the configuration credibility is 70%, if the total number of generated questions does not reach 10, and the number of questions with credibility greater than 70% is less than 5, it means that the configuration condition is not met and question generation continues; otherwise, question generation stops.

[0138] Through the above embodiment, a reasonable number of questions can be automatically generated to avoid wasting time costs. At the same time, since there are enough questions, the requirements for test accuracy are also met.

[0139] S18, determining the risk level of the user according to the credibility corresponding to each question.

[0140] In this embodiment, determining the risk level of the user according to the credibility corresponding to each question includes:

[0141] Acquire, from the generated topics, topics whose credibility is greater than the configured credibility as candidate topics;

[0142] Obtaining the topic with the highest corresponding risk level from the candidate topics as the target topic;

[0143] The risk level corresponding to the target topic is determined as the risk level of the user.

[0144] For example: from the generated questions, the questions with a credibility greater than 70% are obtained as candidate questions, and the question with the highest risk level among the candidate questions is obtained as the target question, and the risk level corresponding to the target question is "aggressive", then the risk level of the user is determined to be "aggressive".

[0145] This embodiment combines micro-expression recognition to determine the credibility of user answers, ensuring that the collected user answers are real feedback from the users, and dynamically adjusts the risk level to make the prediction of the risk level more automated and the prediction results more accurate.

[0146] It should be noted that in order to further improve the security of data and prevent data from being maliciously tampered with, the prediction model can be stored in the blockchain node.

[0147] It can be seen from the above technical solutions that when the present invention detects that a user has triggered an answering instruction, it can obtain the user identity information according to the answering instruction, randomly configure the initial risk level of the user, continuously extract questions from the configured question bank, and generate questions according to the initial risk level and the extracted questions. The generated questions not only include the questions, but also incorporate risk scenarios related to the risk level, making the questions more easy to understand. At the same time, since the generated questions are random in real time, the malicious guidance of relevant personnel is effectively avoided, and the accuracy of the risk assessment results is guaranteed. Each generated question is broadcasted, and the micro-expression features of the user and the answers fed back by the user are collected during the broadcasting process. The features corresponding to each question are constructed according to the initial risk level, the user identity information, the micro-expression features and the answers. The pre-trained prediction model is called, and the features corresponding to each question are input into the prediction model, and the output data of the prediction model is obtained as the credibility corresponding to each question. The initial risk level is adjusted according to the credibility corresponding to each question and the answer corresponding to each question. The credibility predicted by the model and the The user's answer automatically adjusts the initial risk level to make the level determination more accurate, and the total number of generated questions is detected in real time. When the total number and the credibility corresponding to each question reach the configuration conditions, the question generation is stopped, and a reasonable number of questions are automatically generated to avoid wasting time costs. At the same time, since there are a sufficient number of questions, the requirements for test accuracy are also met. Furthermore, the user's risk level is determined according to the credibility corresponding to each question, and the credibility of the user's answer is determined in combination with micro-expression recognition to ensure that the collected user's answers are from the user's real feedback. At the same time, the risk level is dynamically adjusted to make the prediction of the risk level more automated and the prediction result more accurate.

[0148] like Figure 2As shown, it is a functional module diagram of a preferred embodiment of the risk level prediction device based on micro-expression recognition of the present invention. The risk level prediction device 11 based on micro-expression recognition includes an acquisition unit 110, a configuration unit 111, a generation unit 112, a collection unit 113, a construction unit 114, a call unit 115, an adjustment unit 116, a detection unit 117 and a determination unit 118. The module / unit referred to in the present invention refers to a series of computer program segments that can be executed by a processor and can complete fixed functions, which are stored in a memory. In this embodiment, the functions of each module / unit will be described in detail in subsequent embodiments.

[0149] When it is detected that a user triggers an answering instruction, the acquisition unit 110 acquires the user identity information according to the answering instruction.

[0150] In this embodiment, the users may include investors and the like.

[0151] In this embodiment, when it is detected that the user logs into the configuration page or application, it can be determined that the user has triggered the question answering instruction.

[0152] In this embodiment, the user identity information may include, but is not limited to, a combination of one or more of the following information: permanent location, gender, and age.

[0153] The user identity information is used to represent the identity of the user and is non-sensitive information.

[0154] In this embodiment, the acquiring unit 110 acquires the user identity information according to the question answering instruction, including:

[0155] Obtaining login information of the user corresponding to the answer instruction;

[0156] The user identity information is determined according to the login information.

[0157] For example, the user's username can be determined based on the question-answering instruction, and the user's permanent region, gender, age and other information can be further determined based on the username as the user identity information.

[0158] The configuration unit 111 randomly configures the initial risk level of the user.

[0159] The initial risk level may be configured according to actual mission requirements. For example, the initial risk level may include, but is not limited to, aggressive, growth, balanced, stable, and conservative.

[0160] In this embodiment, the initial risk level may be randomly configured for the user, and then dynamically adjusted.

[0161] The generating unit 112 continuously extracts questions from the configured question bank, and generates questions according to the initial risk level and the extracted questions.

[0162] The configuration question bank can be maintained in advance and updated in real time according to actual needs to meet the latest test needs.

[0163] Traditional subjective multiple-choice questions include many financial terms, and the textual expressions are relatively obscure and difficult to understand. In order to complete the questionnaire, investors need to read long paragraphs of text, and the whole process experience is poor.

[0164] For example, a traditional topic could be:

[0165] “When your investments fluctuate to what extent, you will become significantly anxious

[0166]

[0167] Since the questions are fixed, some investors can artificially choose specific answers to distort the evaluation results. Some institutional sales personnel also have the opportunity to induce investors to choose specific answers, thereby making originally unsuitable investors eligible to purchase their recommended products, resulting in inaccurate final risk assessment results.

[0168] In order to solve the above problem, this embodiment can automatically generate questions in real time.

[0169] In this embodiment, the generating unit 112 generates questions according to the initial risk level and the extracted questions, including:

[0170] Acquiring risk scenarios from the configuration question bank according to the initial risk level;

[0171] Get configuration options;

[0172] The extracted question surface, the risk scenario and the configuration options are combined to obtain the question.

[0173] For example, the question is "You have held a stock fund for a year and suddenly made a large provision for losses, resulting in losses", and the initial risk level is "conservative", and the risk scenario corresponding to "conservative" is "loss of principal within 10%", then after combining the question and risk scenario, the question is:

[0174] "You have held a stock fund for a year and suddenly made a large provision for losses, with a 10% loss. Do you show obvious anxiety about this?

[0175] A. Yes

[0176] B. No

[0177] C. It depends on the situation, not necessarily."

[0178] In the above embodiment, the questions are automatically generated by matching the risk level. The generated questions not only include the title, but also incorporate the risk scenarios related to the risk level, making the questions more understandable. At the same time, since the generated questions are random in real time, the malicious guidance of relevant personnel is effectively avoided, ensuring the accuracy of the risk assessment results.

[0179] The collection unit 113 announces each generated question and collects the user's micro-expression features and the answer fed back by the user during the announcement process.

[0180] Specifically, each generated question can be converted into voice, and each generated question can be voice broadcasted to facilitate user evaluation.

[0181] In this embodiment, the collecting unit 113 collects the user's micro-expression features and the answers fed back by the user during the broadcast process, including:

[0182] Invoking a micro-expression recognition model to recognize the micro-expression of the user to obtain the micro-expression feature;

[0183] Starting a preset audio and video acquisition device to collect the voice input by the user;

[0184] The speech input by the user is semantically recognized to obtain an answer fed back by the user.

[0185] The micro-expression recognition model may adopt a facial recognition component module commonly used in the market, and the present invention does not limit the micro-expression recognition model.

[0186] The micro-expression features are used to reflect the user's emotions, as shown in the following table:

[0187]

[0188] The preset audio and video acquisition device may include a camera or a microphone of a terminal that broadcasts the question.

[0189] Among them, a general semantic recognition model can be used to perform semantic recognition on the voice input by the user, and the present invention is not limited thereto.

[0190] The construction unit 114 constructs the features corresponding to each question according to the initial risk level, the user identity information, the micro-expression features and the answers.

[0191] For example: the initial risk level can be configured as R (the values ​​are: aggressive R1, growth R2, balance R3, robust R4, conservative R5), the user identity information can be configured as region D, gender G, age O, the micro-expression features can be configured as T1...Tn, the answer can be configured as A (the values ​​can be yes, no, uncertain), and all the above features can be combined into a feature vector to obtain the features corresponding to each question.

[0192] The calling unit 115 calls a pre-trained prediction model, inputs the features corresponding to each question into the prediction model, and obtains output data of the prediction model as the credibility corresponding to each question.

[0193] In at least one embodiment of the present invention, before calling the pre-trained prediction model, historical answer records are obtained;

[0194] Constructing a training sample according to the historical answer record, wherein the training sample includes at least one question sample, a feature corresponding to each question sample, and a credibility corresponding to each question sample;

[0195] The prediction model is obtained by performing unsupervised learning using the training samples.

[0196] The prediction model may be any network model with prediction function, such as a neural network model, etc., which is not limited by the present invention.

[0197] Through the above implementation, a prediction model can be automatically constructed to achieve automatic prediction of credibility, thereby improving the accuracy of the prediction.

[0198] Furthermore, after the prediction model is trained, the features corresponding to each question can be input into the prediction model, and the output data of the prediction model can be obtained as the credibility corresponding to each question. For example, the credibility corresponding to each question can be a percentage, such as 70%.

[0199] The adjustment unit 116 adjusts the initial risk level according to the credibility corresponding to each question and the answer corresponding to each question.

[0200] Since the initial risk level is obtained by random initialization, in order to obtain an accurate risk level, the initial risk level needs to be adjusted.

[0201] In this embodiment, the adjusting unit 116 adjusts the initial risk level according to the credibility corresponding to each question and the answer corresponding to each question, including:

[0202] Get configuration credibility;

[0203] For each question, when the credibility corresponding to the question is greater than the configured credibility and the answer corresponding to the question is a negative type, reduce the initial risk level according to the configured step size; or

[0204] When the credibility corresponding to the question is greater than the configured credibility, and the answer corresponding to the question is of the affirmative type, the initial risk level is increased according to the configured step size; or

[0205] When the credibility corresponding to the question is greater than the configuration credibility, and the answer corresponding to the question is of an uncertain type, the initial risk level is maintained unchanged; or

[0206] When the credibility corresponding to the question is less than or equal to the configuration credibility, the initial risk level is kept unchanged.

[0207] For example, when the credibility corresponding to the topic is 60% and the configuration credibility is 70%, since the credibility corresponding to the topic is less than the configuration credibility, the initial risk level is kept unchanged.

[0208] In the above embodiment, the initial risk level is automatically adjusted according to the credibility of the model prediction and the user's answer, so that the determination of the risk level is more accurate.

[0209] The detection unit 117 detects the total amount of generated questions in real time, and stops generating questions when the total amount and the credibility corresponding to each question reach the configured conditions.

[0210] It is understandable that when the number of generated questions is too large, it will waste the time of the person being evaluated and increase the time cost of the evaluation; when the number of generated questions is too small, the evaluation results will be inaccurate due to the small number of evaluation questions and insufficient base. Therefore, it is necessary to reasonably determine the number of generated questions.

[0211] Further, obtaining the number of questions whose credibility is greater than the configured credibility from the generated questions as a first number;

[0212] When the total amount is greater than or equal to a first threshold, and / or the first number is greater than or equal to a second threshold, it is determined that the total amount and the credibility corresponding to each question meet the configuration condition.

[0213] The first threshold and the second threshold can be custom configured, for example, the first threshold can be 10, and the second threshold can be 5.

[0214] When the configuration credibility is 70%, if the total number of generated questions does not reach 10, and the number of questions with credibility greater than 70% is less than 5, it means that the configuration condition is not met and question generation continues; otherwise, question generation stops.

[0215] Through the above embodiment, a reasonable number of questions can be automatically generated to avoid wasting time costs. At the same time, since there are enough questions, the requirements for test accuracy are also met.

[0216] The determination unit 118 determines the risk level of the user according to the credibility corresponding to each topic.

[0217] In this embodiment, the determining unit 118 determines the risk level of the user according to the credibility corresponding to each question, including:

[0218] Acquire, from the generated topics, topics whose credibility is greater than the configured credibility as candidate topics;

[0219] Obtaining the topic with the highest corresponding risk level from the candidate topics as the target topic;

[0220] The risk level corresponding to the target topic is determined as the risk level of the user.

[0221] For example: from the generated questions, the questions with a credibility greater than 70% are obtained as candidate questions, and the question with the highest risk level among the candidate questions is obtained as the target question, and the risk level corresponding to the target question is "aggressive", then the risk level of the user is determined to be "aggressive".

[0222] This embodiment combines micro-expression recognition to determine the credibility of user answers, ensuring that the collected user answers are real feedback from the users, and dynamically adjusts the risk level to make the prediction of the risk level more automated and the prediction results more accurate.

[0223] It should be noted that in order to further improve the security of data and prevent data from being maliciously tampered with, the prediction model can be stored in the blockchain node.

[0224] It can be seen from the above technical solutions that when the present invention detects that a user has triggered an answering instruction, it can obtain the user identity information according to the answering instruction, randomly configure the initial risk level of the user, continuously extract questions from the configured question bank, and generate questions according to the initial risk level and the extracted questions. The generated questions not only include the questions, but also incorporate risk scenarios related to the risk level, making the questions more easy to understand. At the same time, since the generated questions are random in real time, the malicious guidance of relevant personnel is effectively avoided, and the accuracy of the risk assessment results is guaranteed. Each generated question is broadcasted, and the micro-expression features of the user and the answers fed back by the user are collected during the broadcasting process. The features corresponding to each question are constructed according to the initial risk level, the user identity information, the micro-expression features and the answers. The pre-trained prediction model is called, and the features corresponding to each question are input into the prediction model, and the output data of the prediction model is obtained as the credibility corresponding to each question. The initial risk level is adjusted according to the credibility corresponding to each question and the answer corresponding to each question. The credibility predicted by the model and the The user's answer automatically adjusts the initial risk level to make the level determination more accurate, and the total number of generated questions is detected in real time. When the total number and the credibility corresponding to each question reach the configuration conditions, the question generation is stopped, and a reasonable number of questions are automatically generated to avoid wasting time costs. At the same time, since there are a sufficient number of questions, the requirements for test accuracy are also met. Furthermore, the user's risk level is determined according to the credibility corresponding to each question, and the credibility of the user's answer is determined in combination with micro-expression recognition to ensure that the collected user's answers are from the user's real feedback. At the same time, the risk level is dynamically adjusted to make the prediction of the risk level more automated and the prediction result more accurate.

[0225] like Figure 3 , which is a schematic diagram of the structure of a computer device of a preferred embodiment of the present invention for implementing a risk level prediction method based on micro-expression recognition.

[0226] The computer device 1 may include a memory 12, a processor 13 and a bus, and may also include a computer program stored in the memory 12 and executable on the processor 13, such as a risk level prediction program based on micro-expression recognition.

[0227] Those skilled in the art will appreciate that the schematic diagram is merely an example of the computer device 1 and does not constitute a limitation on the computer device 1. The computer device 1 may be a bus-type structure or a star-type structure. The computer device 1 may also include more or less other hardware or software than shown in the diagram, or a different arrangement of components. For example, the computer device 1 may also include input and output devices, network access devices, etc.

[0228] It should be noted that the computer device 1 is only an example, and other existing or future electronic products that are suitable for the present invention should also be included in the protection scope of the present invention and included here by reference.

[0229] Among them, the memory 12 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, mobile hard disk, multimedia card, card-type memory (for example: SD or DX memory, etc.), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 12 can be an internal storage unit of the computer device 1, such as a mobile hard disk of the computer device 1. In other embodiments, the memory 12 can also be an external storage device of the computer device 1, such as a plug-in mobile hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (SecureDigital, SD) card, a flash card (Flash Card), etc. equipped on the computer device 1. Further, the memory 12 can also include both an internal storage unit of the computer device 1 and an external storage device. The memory 12 can not only be used to store application software and various types of data installed in the computer device 1, such as the code of the risk level prediction program based on micro-expression recognition, but also can be used to temporarily store data that has been output or is to be output.

[0230] In some embodiments, the processor 13 may be composed of an integrated circuit, for example, a single packaged integrated circuit, or a plurality of packaged integrated circuits with the same or different functions, including one or more central processing units (CPUs), microprocessors, digital processing chips, graphics processors, and combinations of various control chips. The processor 13 is the control core (Control Unit) of the computer device 1, and uses various interfaces and lines to connect the various components of the entire computer device 1, and executes or executes programs or modules stored in the memory 12 (for example, executing a risk level prediction program based on micro-expression recognition, etc.), and calls data stored in the memory 12 to execute various functions of the computer device 1 and process data.

[0231] The processor 13 executes the operating system of the computer device 1 and various installed applications. The processor 13 executes the applications to implement the steps in the above-mentioned various risk level prediction methods based on micro-expression recognition, for example Figure 1 Steps shown.

[0232] Exemplarily, the computer program may be divided into one or more modules / units, which are stored in the memory 12 and executed by the processor 13 to complete the present invention. The one or more modules / units may be a series of computer-readable instruction segments capable of completing specific functions, which are used to describe the execution process of the computer program in the computer device 1. For example, the computer program may be divided into an acquisition unit 110, a configuration unit 111, a generation unit 112, a collection unit 113, a construction unit 114, a call unit 115, an adjustment unit 116, a detection unit 117, and a determination unit 118.

[0233] The above-mentioned integrated unit implemented in the form of a software function module can be stored in a computer-readable storage medium. The above-mentioned software function module is stored in a storage medium, and includes a number of instructions for enabling a computer device (which can be a personal computer, a computer device, or a network device, etc.) or a processor to execute the part of the risk level prediction method based on micro-expression recognition described in various embodiments of the present invention.

[0234] If the module / unit integrated in the computer device 1 is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware devices through a computer program. The computer program can be stored in a computer-readable storage medium, and when the computer program is executed by a processor, the steps of each of the above-mentioned method embodiments can be implemented.

[0235] The computer program includes computer program code, which may be in source code form, object code form, executable file or some intermediate form, etc. The computer readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory, etc.

[0236] Furthermore, the computer-readable storage medium may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function, etc.; the data storage area may store data created according to the use of the blockchain node, etc.

[0237] The blockchain referred to in this invention is a new application mode of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0238] The bus can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 3 The bus is represented by only one straight line, but it does not mean that there is only one bus or one type of bus. The bus is configured to realize the connection and communication between the memory 12 and at least one processor 13, etc.

[0239] Although not shown, the computer device 1 may also include a power source (such as a battery) for supplying power to each component. Preferably, the power source may be logically connected to the at least one processor 13 through a power management device, so that the power management device can realize functions such as charging management, discharging management, and power consumption management. The power source may also include any components such as one or more DC or AC power sources, recharging devices, power failure detection circuits, power converters or inverters, power status indicators, etc. The computer device 1 may also include a variety of sensors, Bluetooth modules, Wi-Fi modules, etc., which will not be repeated here.

[0240] Furthermore, the computer device 1 may also include a network interface. Optionally, the network interface may include a wired interface and / or a wireless interface (such as a WI-FI interface, a Bluetooth interface, etc.), which is usually used to establish a communication connection between the computer device 1 and other computer devices.

[0241] Optionally, the computer device 1 may further include a user interface, which may be a display, an input unit (such as a keyboard), or a standard wired interface or a wireless interface. Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, or an OLED (Organic Light-Emitting Diode) touch device. The display may also be appropriately referred to as a display screen or a display unit, which is used to display information processed in the computer device 1 and to display a visual user interface.

[0242] It should be understood that the embodiment is for illustration only and the scope of the patent application is not limited to this structure.

[0243] Figure 3 Only the computer device 1 having components 12-13 is shown, and it can be understood by those skilled in the art that Figure 3 The structure shown does not constitute a limitation on the computer device 1, and may include fewer or more components than shown in the figure, or combine certain components, or arrange the components differently.

[0244] Combination Figure 1 The memory 12 in the computer device 1 stores a plurality of instructions to implement a risk level prediction method based on micro-expression recognition, and the processor 13 can execute the plurality of instructions to implement:

[0245] When it is detected that a user has triggered an answering instruction, the user identity information is obtained according to the answering instruction;

[0246] Randomly configuring an initial risk level of the user;

[0247] Continuously extracting questions from the configured question bank, and generating questions according to the initial risk level and the extracted questions;

[0248] Announce each generated question, and collect the user's micro-expression features and the user's feedback answer during the announcement process;

[0249] Constructing features corresponding to each question according to the initial risk level, the user identity information, the micro-expression features and the answers;

[0250] Calling a pre-trained prediction model, inputting the features corresponding to each question into the prediction model, and obtaining output data of the prediction model as the credibility corresponding to each question;

[0251] Adjusting the initial risk level according to the credibility of each question and the answer to each question;

[0252] Detect the total amount of generated questions in real time, and stop generating questions when the total amount and the credibility of each question reach the configured conditions;

[0253] The risk level of the user is determined according to the credibility corresponding to each question.

[0254] Specifically, the specific implementation method of the processor 13 for the above instructions can refer to Figure 1 The description of the relevant steps in the corresponding embodiments will not be repeated here.

[0255] It should be noted that the data involved in this case were all obtained legally.

[0256] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only illustrative, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0257] The present invention can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present invention can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present invention can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0258] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0259] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0260] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0261] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.

[0262] In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the present invention can also be implemented by one unit or device through software or hardware. The words first, second, etc. are used to indicate names, and do not indicate any particular order.

[0263] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A risk level prediction method based on micro-expression recognition, characterized in that: The risk level prediction method based on micro-expression recognition includes: When it is detected that a user has triggered an answering instruction, the user identity information is obtained according to the answering instruction; Randomly configuring an initial risk level of the user; Continuously extracting questions from the configured question bank, and generating questions according to the initial risk level and the extracted questions; Announce each generated question, and collect the user's micro-expression features and the user's feedback answer during the announcement process; Constructing features corresponding to each question according to the initial risk level, the user identity information, the micro-expression features and the answers; Calling a pre-trained prediction model, inputting the features corresponding to each question into the prediction model, and obtaining output data of the prediction model as the credibility corresponding to each question; Adjusting the initial risk level according to the credibility of each question and the answer to each question; Detect the total amount of generated questions in real time, and stop generating questions when the total amount and the credibility of each question reach the configured conditions; The risk level of the user is determined according to the credibility corresponding to each question.

2. The risk level prediction method based on micro-expression recognition as claimed in claim 1, characterized in that: The generating of questions according to the initial risk level and the extracted questions includes: Acquiring risk scenarios from the configuration question bank according to the initial risk level; Get configuration options; The extracted question surface, the risk scenario and the configuration options are combined to obtain the question.

3. The risk level prediction method based on micro-expression recognition as claimed in claim 1, characterized in that: The collecting of the user's micro-expression features and the answer fed back by the user during the broadcasting process includes: Invoking a micro-expression recognition model to recognize the micro-expression of the user to obtain the micro-expression feature; Starting a preset audio and video acquisition device to collect the voice input by the user; The speech input by the user is semantically recognized to obtain an answer fed back by the user.

4. The risk level prediction method based on micro-expression recognition as claimed in claim 1, characterized in that: Before calling the pre-trained prediction model, the method further includes: Get historical answer records; Constructing a training sample according to the historical answer record, wherein the training sample includes at least one question sample, a feature corresponding to each question sample, and a credibility corresponding to each question sample; The prediction model is obtained by performing unsupervised learning using the training samples.

5. The risk level prediction method based on micro-expression recognition as claimed in claim 1, characterized in that: The adjusting the initial risk level according to the credibility of each question and the answer to each question includes: Get configuration credibility; For each question, when the credibility corresponding to the question is greater than the configured credibility and the answer corresponding to the question is a negative type, reduce the initial risk level according to the configured step size; or When the credibility corresponding to the question is greater than the configured credibility, and the answer corresponding to the question is of the affirmative type, the initial risk level is increased according to the configured step size; or When the credibility corresponding to the question is greater than the configuration credibility, and the answer corresponding to the question is of an uncertain type, the initial risk level is maintained unchanged; or When the credibility corresponding to the question is less than or equal to the configuration credibility, the initial risk level is kept unchanged.

6. The risk level prediction method based on micro-expression recognition as claimed in claim 5, characterized in that: The method further comprises: Acquire, from the generated topics, the number of topics whose credibility is greater than the configured credibility as a first number; When the total amount is greater than or equal to a first threshold, and / or the first number is greater than or equal to a second threshold, it is determined that the total amount and the credibility corresponding to each question meet the configuration condition.

7. The risk level prediction method based on micro-expression recognition as claimed in claim 1, characterized in that: Determining the risk level of the user according to the credibility corresponding to each question includes: Acquire, from the generated topics, topics whose credibility is greater than the configured credibility as candidate topics; Obtaining the topic with the highest corresponding risk level from the candidate topics as the target topic; The risk level corresponding to the target topic is determined as the risk level of the user.

8. A risk level prediction device based on micro-expression recognition, characterized in that: The risk level prediction device based on micro-expression recognition comprises: An acquisition unit, configured to acquire user identity information according to the answering instruction when detecting that a user has triggered an answering instruction; a configuration unit, configured to randomly configure the initial risk level of the user; A generating unit, configured to continuously extract questions from a configured question bank, and generate questions according to the initial risk level and the extracted questions; A collection unit, used to announce each generated question and collect the user's micro-expression features and the answer fed back by the user during the announcement process; A construction unit, configured to construct a feature corresponding to each question according to the initial risk level, the user identity information, the micro-expression feature and the answer; A calling unit, used to call a pre-trained prediction model, input the features corresponding to each question into the prediction model, and obtain output data of the prediction model as the credibility corresponding to each question; An adjustment unit, configured to adjust the initial risk level according to the credibility corresponding to each question and the answer corresponding to each question; A detection unit, used for detecting the total amount of generated questions in real time, and stopping the generation of questions when the total amount and the credibility corresponding to each question reach a configured condition; The determination unit is used to determine the risk level of the user according to the credibility corresponding to each question.

9. A computer device, characterized in that: The computer device comprises: a memory storing at least one instruction; and A processor executes instructions stored in the memory to implement the risk level prediction method based on micro-expression recognition as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores at least one instruction, and the at least one instruction is executed by a processor in a computer device to implement the risk level prediction method based on micro-expression recognition as described in any one of claims 1 to 7.

Images