Family education evaluation system based on artificial intelligence

Abstract

The invention discloses a family education evaluation system based on artificial intelligence, which comprises a data acquisition module, an answer data acquisition module, an AI behavior analysis module and a dynamic scoring module, and relates to the technical field of family education. The family education quality score is obtained by adding the learning time score of the child, the time score spent by the parents in learning of the child and the answer test accuracy score, the limitation of a single index is broken through, and the education quality is reflected more comprehensively; a learning state judgment model is constructed through a convolutional neural network algorithm, whether a child is in a learning state is judged, whether parents join child learning is judged, automatic and objective monitoring is achieved, and the influence of subjective factors is avoided; a piecewise function scoring method is adopted, scores are dynamically distributed according to a time threshold value generated by K-means clustering, and limitation of fixed scores and subjective setting deviation are avoided; the technical problems that traditional family education evaluation is high in subjectivity and fixed in weight are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of family education, and in particular to a family education evaluation system based on artificial intelligence. Background Art

[0002] Family education is one of the most important parts of a child's education system and is also the foundation of school education and social education. Good family education can ensure the healthy physical and mental development of children before they enter society and receive collective education, and also lay a good foundation for receiving education in kindergartens and schools.

[0003] Publication number CN115456826B discloses a method and system for evaluating the excellence index of a family education model. By having parents and children complete test questions at the same time and making a comprehensive judgment based on the scores of the questions, the family education model can be comprehensively evaluated from the different perspectives of parents and children to ensure the accuracy of the evaluation. By collecting the answering status characteristics of parents and children, the scoring status of parents and children is analyzed based on the answering status characteristics to determine whether the users have completed the test questions carefully, thereby ensuring the effectiveness and accuracy of the test questions completed by parents and children.

[0004] However, the above application still has the following problems: when conducting family education assessments, the method used is to have parents and children complete the test questions at the same time, which has limitations. On the one hand, this method is too dependent on the quality and pertinence of the test questions themselves. On the other hand, this method mainly evaluates the performance of parents and children during the answering process, which is very subjective. The state and mood of children and parents when answering the questions may affect the answering results, making the evaluation results less objective. Moreover, this method only focuses on the performance in the specific scenario of answering questions, and does not refer to the child’s learning time or the time parents spend participating in their children’s learning, and cannot provide a more comprehensive understanding of the child’s daily learning. Summary of the Invention

[0005] In view of the shortcomings of the existing technology, the present invention proposes a family education assessment system based on artificial intelligence.

[0006] An artificial intelligence-based family education assessment system, comprising:

[0007] Data collection module: uses cameras and microphones to collect video and voice data of children's learning scenes in real time;

[0008] Question answering data collection module: Generate at least one question answering test in a unit of time through the online education platform or smart device, and obtain the child's answer data and the accuracy rate of the question answering test;

[0009] AI behavior analysis module: This module obtains video and voice data from the data acquisition module and uses a convolutional neural network algorithm to build a learning status judgment model to determine whether the child is in a learning state and whether the parents are participating in the child's learning.

[0010] If the child is judged to be in a learning state, record the child's learning time;

[0011] If it is determined that the parent is participating in the child's study, the time the parent spends participating in the child's study is recorded;

[0012] Dynamic scoring module: used to calculate family education quality scores.

[0013] Preferably, in the dynamic scoring module, the family education quality score is calculated as follows:

[0014] In unit time, obtain the child's learning time t1, the time t2 spent by parents joining their children's learning, and the answer test accuracy A in the AI ​​behavior analysis module;

[0015] According to the preset time thresholds of t1 and t2, the score W2 of t1 and the score W1 of t2 are calculated by the piecewise function scoring method;

[0016] Then the score W3 of A is: W3 = total score - W1 - W2;

[0017] By calculating the ratio of t1 or t2 to the preset time threshold and multiplying it by W2 or W1, the score of t1 or t2 is obtained;

[0018] By multiplying A by W3, we can get the accuracy score of the test.

[0019] Add the scores of t1, t2 and A to get the family education quality score.

[0020] Preferably, in the dynamic scoring module, the preset time thresholds of t1 and t2 are obtained by the following method:

[0021] The AI ​​behavior analysis module is used to obtain historical data on children's learning time and the time parents spent on their children's learning.

[0022] Use the K-means algorithm to cluster the historical children's learning time data, and divide the historical children's learning time data into three clusters: low, medium, and high. The cluster centers of the low, medium, and high clusters are used as the time threshold of t1;

[0023] The K-means algorithm is used to cluster the historical data on the time parents spent on their children's learning, and the historical data on the time parents spent on their children's learning are divided into three clusters: low, medium, and high. The cluster centers of the low, medium, and high clusters are used as the time thresholds of t2.

[0024] 5. Preferably, it also includes: a scoring module: used to score the quality of family education, as follows:

[0025] In unit time, the learning time of the child obtained from the AI ​​behavior analysis module is t1. t1 is the sum of multiple continuous learning periods. Let the number of continuous learning periods be n. The continuous learning periods and their number n are obtained through artificial intelligence algorithm analysis. n is a positive integer. When n is less than or equal to the set threshold, no points are assigned. When n is greater than the set threshold, the following points are assigned:

[0026] Let the time threshold be t b ;

[0027] When the continuous learning time is greater than or equal to the threshold t b The number is greater than When , add m points to the result of family education quality score p;

[0028] When the continuous learning time is greater than or equal to the threshold t b The number is greater than When , add v points to the result of family education quality score p;

[0029] When the continuous learning time is greater than or equal to the threshold t b The number is greater than When , add u points to the result of family education quality score p;

[0030] And 0<m<v<u.

[0031] A family education assessment method based on artificial intelligence, comprising the following steps:

[0032] S1, collects video data and voice data of children’s learning scenes in real time through cameras and microphones;

[0033] S2. Generate at least one test within a given time period through an online education platform or smart device, and obtain the child's answer data and test accuracy.

[0034] S3, obtain the video data and voice data in S1, build a learning status judgment model through the convolutional neural network algorithm, judge whether the child is in the learning state, and judge whether the parents join the child's learning;

[0035] If the child is judged to be in a learning state, record the child's learning time;

[0036] If it is determined that the parent is participating in the child's study, the time the parent spends participating in the child's study is recorded;

[0037] S4. Calculate the family education quality score as follows:

[0038] In unit time, obtain the child’s study time t1, the time t2 spent by the parent in joining the child’s study, and the answer test accuracy A in S3;

[0039] According to the preset time thresholds of t1 and t2, the score W2 of t1 and the score W1 of t2 are calculated by the piecewise function scoring method;

[0040] Then the score W3 of A is: W3 = total score - W1 - W2;

[0041] By calculating the ratio of t1 or t2 to the preset time threshold and multiplying it by W2 or W1, the score of t1 or t2 is obtained;

[0042] By multiplying A by W3, we can get the accuracy score of the test.

[0043] Add the scores of t1, t2 and A to get the family education quality score.

[0044] Preferably, let the family education quality score be p, the total score of p is 100, if p is less than 60, an alarm is issued;

[0045] If 60≤p≤80, issue an early warning;

[0046] If 80<p≤100, no early warning or alarm will be issued.

[0047] Preferably, the accuracy rate of answering the test within a unit time is recorded as A;

[0048] If A is less than 60%, all questions in the test are marked and a set of wrong questions is generated;

[0049] If 60%≤A≤80%, mark the wrong questions in the test and generate a wrong question set;

[0050] If 80% < A ≤ 100%, all questions in the test are not marked.

[0051] Preferably, if p<60, the questionnaire is pushed and the questionnaire is a text questionnaire;

[0052] If 60≤p≤80, push the questionnaire, and the questionnaire is a voice questionnaire;

[0053] If 80<p≤100, the questionnaire will not be pushed.

[0054] A terminal includes a processor and a storage medium; the storage medium is used to store instructions;

[0055] The processor is used to operate according to the instructions to execute the steps of the above-mentioned artificial intelligence-based family education assessment system.

[0056] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-mentioned artificial intelligence-based family education assessment system.

[0057] The artificial intelligence-based family education assessment system proposed in this invention has the following beneficial technical effects:

[0058] 1. By adding the child's learning time score, the parent's time spent participating in the child's learning score, and the test accuracy score, we get the family education quality score, breaking through the limitations of a single indicator and reflecting the quality of education more comprehensively;

[0059] Through cameras and microphones, real-time video and voice data of children's learning scenes are collected. A learning status judgment model is constructed using a convolutional neural network algorithm to determine whether the child is in a learning state and whether the parents are participating in the child's learning. This enables automated and objective monitoring and avoids the influence of subjective factors.

[0060] Using a piecewise function scoring method, we dynamically assign scores to children's learning time and the time parents spend participating in their children's learning based on the time threshold generated by K-means clustering, avoiding the limitations of fixed scores. The time threshold generated by K-means clustering is generated through clustering of historical data to avoid subjective setting bias.

[0061] By calculating the ratio of the child's study time or the time parents spend participating in their child's study to the preset time threshold, and using the ratio of actual data to the threshold as a weighting factor, the differences can be further quantified within the same threshold range, reflecting the quality of family education more accurately and solving the technical problems of traditional family education assessment, which is highly subjective and has fixed weights.

[0062] 2. Setting of the scoring module: per unit time, the learning time of the child in the AI ​​behavior analysis module is t1, t1 is the sum of multiple continuous learning times, and the number of continuous learning times is n. The continuous learning time and its number n are obtained through artificial intelligence algorithm analysis. When n is greater than the set threshold, points are assigned according to the number of continuous learning times greater than or equal to the time threshold, and the bonus value increases according to the number of continuous learning times greater than or equal to the time threshold. This step-by-step scoring mechanism can motivate children to maintain continuous learning for a longer period of time, cultivate good learning habits, and help improve concentration. It has a positive effect on cultivating children's learning perseverance and concentration, and improves the accuracy of family education quality scores. In addition, this data-based guidance method is more scientific and targeted, which helps to improve the quality of family education.

[0063] Additional aspects and advantages of the present invention will be set forth in part in the description which follows and, in part, will be obvious from the description which follows, or may be learned by practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0064] Figure 1 It is a principle block diagram of the system of the present invention;

[0065] Figure 2 Flowchart of the method of the present invention. DETAILED DESCRIPTION

[0066] The following describes embodiments of the present invention in detail. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar symbols throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended only to explain the present invention, and are not to be construed as limiting the present invention.

[0067] like Figure 1 The artificial intelligence-based family education assessment system shown includes:

[0068] Data collection module: uses cameras and microphones to collect video and voice data of children's learning scenes in real time;

[0069] Question answering data collection module: Generate at least one question answering test in a unit of time through the online education platform or smart device, and obtain the child's answer data and the accuracy rate of the question answering test;

[0070] Smart devices include learning tablets and mobile phones;

[0071] When using this application, data encryption technology is used to encrypt data during transmission and storage to ensure data security;

[0072] AI behavior analysis module: This module obtains video and voice data from the data acquisition module and uses a convolutional neural network algorithm to build a learning status judgment model to determine whether the child is in a learning state and whether the parents are participating in the child's learning.

[0073] If the child is judged to be in a learning state, record the child's learning time;

[0074] If it is determined that the parent is participating in the child's study, the time the parent spends participating in the child's study is recorded;

[0075] Dynamic scoring module: used to calculate the quality score of family education;

[0076] In an optional embodiment, in the dynamic scoring module, the family education quality score is calculated as follows:

[0077] In unit time, obtain the child's learning time t1, the time t2 spent by parents joining their children's learning, and the answer test accuracy A in the AI ​​behavior analysis module;

[0078] According to the preset time thresholds of t1 and t2, the score W2 of t1 and the score W1 of t2 are calculated by the piecewise function scoring method;

[0079] Then the score W3 of A is: W3 = total score - W1 - W2;

[0080] By calculating the ratio of t1 or t2 to the preset time threshold and multiplying it by W2 or W1, the score of t1 or t2 is obtained;

[0081] By multiplying A by W3, we can get the accuracy score of the test.

[0082] The scores of t1, t2, and A are added together to obtain the family education quality score;

[0083] In an optional embodiment, in the dynamic scoring module, the preset time thresholds t1 and t2 are obtained by the following method:

[0084] The AI ​​behavior analysis module is used to obtain historical data on children's learning time and the time parents spent on their children's learning.

[0085] Use the K-means algorithm to cluster the historical children's learning time data, and divide the historical children's learning time data into three clusters: low, medium, and high. The cluster centers of the low, medium, and high clusters are used as the time threshold of t1;

[0086] Use the K-means algorithm to cluster the historical data on the time parents spent on their children's learning, and divide the historical data into three clusters: low, medium, and high. The cluster centers of the low, medium, and high clusters are used as the time threshold of t2;

[0087] In an optional embodiment, in the dynamic scoring module, the family education quality score is calculated as follows:

[0088] Let the total score be 100;

[0089] In unit time, record the child’s study time as t1, set the score of the child’s study time as W2, and set the time threshold as t a , scored as:

[0090] Set the time threshold t a t a1 , t a2 , t a3 , and 0<t a1 <t a2 <t a3 ;

[0091] If t1≤t a1 , then W2=45; the score is:

[0092] If t a1 <t1≤t a2 , then W2=50; the score is:

[0093] If t a2 <t1≤t a3 , then W2=55; the score is:

[0094] If t a3 <t1, then W2=60, and the score is:

[0095] In unit time, the time spent by parents in joining their children’s learning is recorded as t2, which accounts for W1. The time threshold is set to t s , scored as:

[0096] Set the time threshold t s t s1 , t s2 , t s3 , and 0<t s1 <t s2 <t s3 ;

[0097] If t2≤t s1 , then W1=12.5; the score is:

[0098] If t s1 <t2≤t s2 , then W1=15; the score is:

[0099] If t s2 <t2≤t s3 , then W1=17.5; the score is:

[0100] If t s3 <t2, then W1=20; the score is:

[0101] In unit time, record the accuracy of the test as A. Assume that the accuracy of the test is W3, and the score is: A×W3, W3=100-W1-W2;

[0102] Assume that the family education quality score is p, then:

[0103]

[0104] By adding the child's learning time score, the parent's time spent participating in the child's learning score, and the test accuracy score, we can get the family education quality score, breaking through the limitations of a single indicator and reflecting the quality of education more comprehensively.

[0105] Through cameras and microphones, real-time video and voice data of children's learning scenes are collected. A learning status judgment model is constructed using a convolutional neural network algorithm to determine whether the child is in a learning state and whether the parents are participating in the child's learning. This enables automated and objective monitoring and avoids the influence of subjective factors.

[0106] Using a piecewise function scoring method, we dynamically assign scores to children's learning time and the time parents spend participating in their children's learning based on the time threshold generated by K-means clustering, avoiding the limitations of fixed scores. The time threshold generated by K-means clustering is generated through clustering of historical data to avoid subjective setting bias.

[0107] By calculating the ratio of the child's study time or the time parents spend participating in their child's study to the preset time threshold, and using the ratio of actual data to the threshold as a weighting factor, the differences can be further quantified within the same threshold range, reflecting the quality of family education more accurately and solving the technical problems of traditional family education assessment, which is highly subjective and has fixed weights.

[0108] Scoring module: used to assign scores to the quality of family education;

[0109] In unit time, the learning time of the child obtained from the AI ​​behavior analysis module is t1. t1 is the sum of multiple continuous learning periods. Let the number of continuous learning periods be n. The continuous learning periods and their number n are obtained through artificial intelligence algorithm analysis. n is a positive integer. When n≤3, no points are assigned. When n>3, the following points are assigned:

[0110] Let the time threshold be t b ;

[0111] When the continuous learning time is greater than or equal to the threshold t b The number is greater than When , add m points to the result of family education quality score p;

[0112] When the continuous learning time is greater than or equal to the threshold t b The number is greater than When , add v points to the result of family education quality score p;

[0113] When the continuous learning time is greater than or equal to the threshold t b The number is greater than When , add u points to the result of family education quality score p;

[0114] And 0<m<v<u.

[0115] Continuous learning time refers to the length of time a child continues to study without interruption, which can reflect the child's sustained concentration;

[0116] In an optional embodiment, m=5; v=10; u=15.

[0117] The purpose of this design is to assign points to the quality of family education based on the continuous learning time.

[0118] The incentive mechanism for continuous study time further strengthens children's focus. When the number of consecutive study hours exceeds or equals different thresholds and reaches a certain level, points are added to the family education quality score p, and the bonus values ​​m, v, and u increase in sequence. This encourages children to maintain continuous study for as long as possible and improve their sustained focus. Children who reach higher continuous study time standards can receive more bonus points, which has a positive effect on cultivating children's learning perseverance and focus.

[0119] The setting of the scoring module is that per unit time, the learning time of the child in the AI ​​behavior analysis module is obtained as t1, t1 is formed by the sum of multiple continuous learning times, and the number of continuous learning times is set to n. The continuous learning time and its number n are obtained through artificial intelligence algorithm analysis. When n is greater than the set threshold, points are assigned according to the number of continuous learning times greater than or equal to the time threshold, and the bonus value increases according to the number of continuous learning times greater than or equal to the time threshold. This step-by-step scoring mechanism can motivate children to maintain continuous learning for a longer period of time, cultivate good learning habits, and help improve concentration. It has a positive effect on cultivating children's learning perseverance and concentration, and improves the accuracy of family education quality scores. In addition, this data-based guidance method is more scientific and targeted, which helps to improve the quality of family education.

[0120] like Figure 1 The artificial intelligence-based family education assessment method shown includes the following steps:

[0121] S1, collects video data and voice data of children’s learning scenes in real time through cameras and microphones;

[0122] S2. Generate at least one test within a given time period through an online education platform or smart device, and obtain the child's answer data and test accuracy.

[0123] The accuracy rate of the test can reflect the child's knowledge mastery. The accuracy rate of the test here refers to the comprehensive accuracy rate of one or more test responses. For example, if three test responses are generated, the accuracy rate is the sum of the accuracy rates of the three test responses divided by three.

[0124] S3, obtain the video data and voice data in S1, build a learning status judgment model through the convolutional neural network algorithm, judge whether the child is in the learning state, and judge whether the parents join the child's learning;

[0125] If the child is judged to be in a learning state, record the child's learning time;

[0126] If it is determined that the parent is participating in the child's study, the time the parent spends participating in the child's study is recorded;

[0127] S4. Calculate the family education quality score as follows:

[0128] In unit time, obtain the child’s study time t1, the time t2 spent by the parent in joining the child’s study, and the answer test accuracy A in S3;

[0129] According to the preset time thresholds of t1 and t2, the score W2 of t1 and the score W1 of t2 are calculated by the piecewise function scoring method;

[0130] Then the score W3 of A is: W3 = total score - W1 - W2;

[0131] By calculating the ratio of t1 or t2 to the preset time threshold and multiplying it by W2 or W1, the score of t1 or t2 is obtained;

[0132] By multiplying A by W3, we can get the accuracy score of the test.

[0133] The scores of t1, t2, and A are added together to obtain the family education quality score;

[0134] Children's learning time data includes the time spent on learning knowledge and answering tests;

[0135] This method can more comprehensively evaluate the level of family participation in education and the child's learning status by recording the child's learning time per unit time, the time parents spend participating in the child's learning, and the accuracy of the child's test answers.

[0136] Based on the preset time thresholds t1 and t2, the piecewise function scoring method is used to calculate the score W2 of t1 and the score W1 of t2. That is, the values ​​of W1 and W2 are automatically adjusted according to the set time thresholds reached by t2 and t1. The recording of parental participation time helps measure the degree of family attention to children's education and provides a basis for promoting parent-child education. This design can motivate parents to participate more in their children's learning, while also prompting children to increase their study time, encouraging children to develop good study habits.

[0137] In an optional embodiment, the unit time can be set to one week, and the time threshold t s1 =7h,t s2 =14h,t s3 =21h;t a1 =16h,t a2 =27h,t a3 =36h; h refers to hours;

[0138] Let the family education quality score be p, the total score of p is 100, if p is less than 60, an alarm is issued;

[0139] If 60≤p≤80, issue an early warning;

[0140] If 80<p≤100, no warning or alarm will be issued;

[0141] The early warning mechanism can promptly remind parents to pay attention to their children's learning status. If p < 60, an alarm is triggered; if 60 ≤ p ≤ 80, an early warning is issued. This allows parents to take timely measures when their children's learning problems arise, adjust their teaching methods and strategies, and improve the effectiveness of family education.

[0142] Let the family education quality score be p. If p < 60, the questionnaire will be pushed, and the questionnaire is a text questionnaire.

[0143] If 60≤p≤80, push the questionnaire, and the questionnaire is a voice questionnaire;

[0144] If 80<p≤100, no questionnaire push will be conducted;

[0145] Depending on the range of the family education quality score p, different questionnaires will be pushed or not pushed. When p < 60, a text questionnaire will be pushed to help parents and children deeply analyze learning problems; when 60 ≤ p ≤ 80, a voice questionnaire will be pushed to more conveniently obtain feedback information. This personalized questionnaire push method can provide targeted guidance and suggestions based on the child's learning situation.

[0146] The value per unit time is greater than t1 and the time threshold t a The value of

[0147] The value of the unit time is greater than t2 and the time threshold t s The value of

[0148] Record the accuracy of the test in unit time, A;

[0149] If A is less than 60%, all questions in the test are marked and a set of wrong questions is generated;

[0150] If 60%≤A≤80%, mark the wrong questions in the test and generate a wrong question set;

[0151] If 80% < A ≤ 100%, all questions in the test will not be marked;

[0152] You can also manually mark any question in the test and include it in the wrong question collection.

[0153] Recording the accuracy of test answers not only reflects a child's knowledge mastery, but also generates collections of incorrect questions based on different accuracy ranges, facilitating targeted review and knowledge consolidation. For example, when the test accuracy is below 60%, all questions are marked to create a collection of incorrect questions, helping the child comprehensively identify and address gaps. When the accuracy is between 60% and 80%, only incorrect questions are marked, improving the efficiency of error collection. When the accuracy is above 80%, no questions are marked, but any question can be manually marked for inclusion in the collection, meeting personalized learning needs.

[0154] At the same time, the contents not described in detail in this specification belong to the existing technology well known to those skilled in the art.

[0155] In the embodiments provided by the present invention, it should be understood that the disclosed systems or methods can be implemented in other ways. For example, the embodiments of the invention described above are merely illustrative. For example, the division of modules is only a logical function division, and other division methods may be used in actual implementation.

[0156] Modules described as separate components may or may not be physically separate, and components shown as modules may or may not be physical modules, and may be located in one place or distributed across multiple network modules. Some or all of these modules may be selected to achieve the objectives of this embodiment based on actual needs.

[0157] In addition, the functional modules in various embodiments of the present invention may be integrated into a single processing module, each module may exist physically separately, or two or more modules may be integrated into a single module. The aforementioned integrated modules may be implemented in the form of hardware or hardware plus software functional modules.

[0158] 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 basic characteristics of the present invention.

[0159] The above description is only a preferred specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with the technical field, within the technical scope disclosed by the present invention, who makes equivalent replacements or changes based on the technical solution and inventive concept of the present invention, should be covered by the scope of protection of the present invention.

Claims

1. A family education assessment system based on artificial intelligence, characterized by: include: Data collection module: uses cameras and microphones to collect video and voice data of children's learning scenes in real time; Question answering data collection module: Generate at least one question answering test in a unit of time through the online education platform or smart device, and obtain the child's answer data and the accuracy rate of the question answering test; AI behavior analysis module: This module obtains video and voice data from the data acquisition module and uses a convolutional neural network algorithm to build a learning status judgment model to determine whether the child is in a learning state and whether the parents are participating in the child's learning. If the child is judged to be in a learning state, record the child's learning time; If it is determined that the parent is participating in the child's study, the time the parent spends participating in the child's study is recorded; Dynamic scoring module: used to calculate family education quality scores.

2. The artificial intelligence-based family education assessment system according to claim 1 is characterized in that: In the dynamic scoring module, the family education quality score is calculated as follows: In unit time, obtain the child's learning time t1, the time t2 spent by parents joining their children's learning, and the answer test accuracy A in the AI ​​behavior analysis module; According to the preset time thresholds of t1 and t2, the score W2 of t1 and the score W1 of t2 are calculated by the piecewise function scoring method; Then the score W3 of A is: W3 = total score - W1 - W2; By calculating the ratio of t1 or t2 to the preset time threshold and multiplying it by W2 or W1, the score of t1 or t2 is obtained; By multiplying A by W3, we can get the accuracy score of the test. Add the scores of t1, t2 and A to get the family education quality score.

3. The artificial intelligence-based family education assessment system according to claim 2 is characterized in that: In the dynamic scoring module, the preset time thresholds of t1 and t2 are obtained by the following method: The AI ​​behavior analysis module is used to obtain historical data on children's learning time and the time parents spent on their children's learning. Use the K-means algorithm to cluster the historical children's learning time data, and divide the historical children's learning time data into three clusters: low, medium, and high. The cluster centers of the low, medium, and high clusters are used as the time threshold of t1; The K-means algorithm is used to cluster the historical data on the time parents spent on their children's learning, and the historical data on the time parents spent on their children's learning are divided into three clusters: low, medium, and high. The cluster centers of the low, medium, and high clusters are used as the time thresholds of t2.

4. The artificial intelligence-based family education assessment system according to claim 3 is characterized in that: Also includes: Scoring module: used to assign scores to family education quality, as follows: In unit time, the learning time of the child obtained from the AI ​​behavior analysis module is t1. t1 is the sum of multiple continuous learning periods. Let the number of continuous learning periods be n. The continuous learning periods and their number n are obtained through artificial intelligence algorithm analysis. n is a positive integer. When n is less than or equal to the set threshold, no points are assigned. When n is greater than the set threshold, the following points are assigned: Let the time threshold be t b ; When the continuous learning time is greater than or equal to the threshold t b The number is greater than When , add m points to the result of family education quality score p; When the continuous learning time is greater than or equal to the threshold t b The number is greater than When , add v points to the result of family education quality score p; When the continuous learning time is greater than or equal to the threshold t b The number is greater than When , add u points to the result of family education quality score p; And 0<m<v<u.

5. The artificial intelligence-based family education assessment method according to any one of claims 1 to 4, characterized in that: The following steps are involved: S1, collects video data and voice data of children’s learning scenes in real time through cameras and microphones; S2. Generate at least one test within a given time period through an online education platform or smart device, and obtain the child's answer data and test accuracy. S3, obtain the video data and voice data in S1, build a learning status judgment model through the convolutional neural network algorithm, judge whether the child is in the learning state, and judge whether the parents join the child's learning; If the child is judged to be in a learning state, record the child's learning time; If it is determined that the parent is participating in the child's study, the time the parent spends participating in the child's study is recorded; S4. Calculate the family education quality score as follows: In unit time, obtain the child’s study time t1, the time t2 spent by the parent in joining the child’s study, and the answer test accuracy A in S3; According to the preset time thresholds of t1 and t2, the score W2 of t1 and the score W1 of t2 are calculated by the piecewise function scoring method; Then the score W3 of A is: W3 = total score - W1 - W2; By calculating the ratio of t1 or t2 to the preset time threshold and multiplying it by W2 or W1, the score of t1 or t2 is obtained; By multiplying A by W3, we can get the accuracy score of the test. Add the scores of t1, t2 and A to get the family education quality score.

6. The artificial intelligence-based family education assessment method according to claim 5, characterized in that: Let the family education quality score be p, the total score of p is 100, if p is less than 60, an alarm is issued; If 60≤p≤80, issue an early warning; If 80<p≤100, no early warning or alarm will be issued.

7. The family education assessment method based on artificial intelligence according to claim 5 is characterized in that: Record the accuracy of the test in unit time, A; If A is less than 60%, all questions in the test are marked and a set of wrong questions is generated; If 60%≤A≤80%, mark the wrong questions in the test and generate a wrong question set; If 80% < A ≤ 100%, all questions in the test are not marked.

8. The family education assessment method based on artificial intelligence according to claim 6 is characterized in that: If p<60, the questionnaire will be pushed and the questionnaire will be a text questionnaire; If 60≤p≤80, push the questionnaire, and the questionnaire is a voice questionnaire; If 80<p≤100, the questionnaire will not be pushed.

9. A terminal comprising a processor and a storage medium; characterized in that: The storage medium is used to store instructions; The processor is used to operate according to instructions to execute the steps of the artificial intelligence-based family education assessment system according to any one of claims 1-4.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the artificial intelligence-based family education assessment system described in any one of claims 1 to 4 are implemented.

Images