A classroom teaching effect evaluation system

Abstract

The present application relates to the field of classroom evaluation, especially to a classroom teaching effect evaluation system, the present application sets data storage module, know and do mapping module, know and do interaction module, correlation control module and wisdom evaluation module, determine the behavior effective area, and the student reaction data and the language data of the teacher in the area are analyzed in advance, to determine the key teaching frame, build the behavior time domain curve and the language time domain curve of the predetermined time period, compare with the reaction time domain curve, locate the abnormal characteristics, call the corresponding module, correlation control module generates the correction characteristics, and the wisdom evaluation module evaluates the teaching effect. The present application quantifies the intelligent classroom cognitive activity, solves the problem that the existing technology analysis cannot cover important dimensions such as classroom thinking mode and teaching style, determines the know and do mapping state through the language data of the teacher and the student reaction data, further analyzes the abnormal characteristics combined with the behavior data, improves the evaluation efficiency and accuracy.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of classroom evaluation, in particular to a classroom teaching effect evaluation system. BACKGROUND

[0002] Classroom teaching evaluation is a process of collecting evidence by means of observation, recording and informationization tools and means, judging, digging and improving the value of classroom teaching activities according to evaluation standards. For a long time, because of the complexity of classroom teaching, classroom evaluation mostly adopts manual recording method. In recent years, with the development and maturity of new generation information technologies such as cloud computing, big data and artificial intelligence, classroom teaching evaluation has the technical conditions of normalization and scaling, and evaluators start to collect semi-structured and unstructured data by using video, audio, education software and education equipment, and the tools, models, theories and practices for observing and evaluating classroom teaching are continuously developed and enriched.

[0003] Chinese patent publication No. CN118782096A discloses a classroom teaching effect evaluation system based on voice multi-feature progressive embedding, which includes an input module, a preprocessing module, a role classification module, a classroom content emotion classification module and an evaluation analysis module. The input module obtains the classroom audio to be evaluated and analyzed. The preprocessing module performs fine-grained cutting on the audio. The role classification module divides each segment into a teacher or a student after detecting and marking the silent segment. The classroom content emotion classification module classifies the teacher segment and the student segment. The evaluation analysis module inputs the proportion information of each category into the classroom evaluation module to generate an evaluation analysis report. The present application can automatically, objectively and efficiently evaluate the classroom teaching effect through progressive voice analysis and optimized teaching evaluation system, thereby improving the teaching quality of teachers.

[0004] Chinese patent publication No. CN118761880A discloses a classroom teaching quality evaluation method based on a large language model, which mainly includes: based on classroom monitoring video data and teaching management system data, using artificial intelligence technology to extract voice, image and text multi-modal features of teachers, students and classroom environment; constructing a course knowledge graph according to a course plan and course information; enhancing the large language model through the constructed course knowledge graph, and realizing intelligent classroom teaching quality evaluation based on the enhanced large language model; the beneficial effects of the present application are: the large language model is enhanced by classroom teaching quality evaluation knowledge and course knowledge graph, and the large language model is guided to generate evaluation result description conforming to classroom teaching quality evaluation fact according to the semantic association of classroom teaching quality evaluation knowledge and course knowledge graph; thereby the model can generate more reasonable and more conforming description to classroom teaching quality evaluation fact.

[0005] It can be seen that the prior art still has the following problems,

[0006] The existing classroom teaching evaluation mainly adopts subjective classroom observation, student questionnaire survey or teacher self-report questionnaire survey, etc. There are serious empiricism, lack of appropriate theory and other tendencies, which makes the evaluation difficult to associate the teaching process and effect, lacks pertinence, leads to evaluation deviation, reduces analysis efficiency and evaluation accuracy. SUMMARY

[0007] Therefore, the present application provides a classroom teaching effect evaluation system to overcome the problems of the existing classroom teaching evaluation mainly adopting subjective classroom observation, student questionnaire survey or teacher self-report questionnaire survey, etc. There are serious empiricism, lack of appropriate theory and other tendencies, which makes the evaluation difficult to associate the teaching process and effect, lacks pertinence, leads to evaluation deviation, reduces analysis efficiency and evaluation accuracy.

[0008] To achieve the above purpose, the present application provides a classroom teaching effect evaluation system, comprising,

[0009] A data storage module, comprising a teacher end for obtaining behavior data and language data of a teacher to determine a behavior effective area, and a student end for obtaining reaction data of a student to determine a knowledge-action delay and a knowledge-action matching degree of the student;

[0010] A knowledge-action mapping module connected with the data storage module, for calculating a knowledge-action characteristic value based on the knowledge-action delay and the knowledge-action matching degree in the behavior effective area, calculating a language characteristic value based on the language data, determining a knowledge-action mapping characteristic value, analyzing a knowledge-action mapping state, and determining a key teaching frame;

[0011] A knowledge-action interaction module connected with the knowledge-action mapping module, for extracting behavior data, language data and reaction data in a predetermined time period, comparing behavior time domain curves, language time domain curves and reaction time domain curves, positioning abnormal characteristics, calculating abnormal characteristic coefficients, and determining a module to be called;

[0012] An association control module connected with the knowledge-action interaction module, for generating a correction feature based on a behavior language database, and sending the correction feature to a storage unit associated with the teacher end;

[0013] A smart evaluation module connected with the knowledge-action interaction module, for storing a teaching video segment corresponding to the predetermined time period to a teaching effect evaluation database, and evaluating a classroom teaching effect;

[0014] The behavior data includes gesture amplitude and gesture trajectory, and the reaction data includes reaction time and visual target.

[0015] Further, the data storage module determines a behavior effective area, comprising,

[0016] determining a pointing position of the behavior data, and determining a specified region containing a minimum region of the pointing position;

[0017] identifying a gesture trajectory of the behavior data, and determining a closed region enclosed by the gesture trajectory;

[0018] determining the specified region and / or the closed region as a behavior effective region.

[0019] Further, the data storage module determines a knowledge-action delay and a knowledge-action matching degree of the student, including,

[0020] determining an absolute value of a difference between the reaction time and the end time of the language data as the knowledge-action delay;

[0021] determining a ratio between a total number of the line-of-sight targets and a number of abnormal line-of-sight targets as the knowledge-action matching degree;

[0022] wherein the number of abnormal line-of-sight targets is a number of line-of-sight targets different from a predetermined proportion of line-of-sight targets.

[0023] Further, the knowledge-action mapping module calculates a knowledge-action feature value, including,

[0024] determining a ratio between the knowledge-action delay and a reference knowledge-action delay as a delay influence factor;

[0025] determining a ratio between a reference knowledge-action matching degree and the knowledge-action matching degree as a matching influence factor;

[0026] determining an average value of a sum of reciprocals of the delay influence factor and the matching influence factor as the knowledge-action feature value.

[0027] Further, the knowledge-action mapping module calculates a language feature value, including,

[0028] identifying non-standard expressions in the language data, and determining a ratio between a number of the non-standard expressions and a total number of expressions as a non-standard expression ratio;

[0029] determining a logical fluency of the language data;

[0030] determining an average value of a sum of reciprocals of the non-standard expression ratio and the logical fluency as the language feature value.

[0031] Further, the knowledge-action mapping module determines a knowledge-action mapping feature value, including,

[0032] determining a ratio between the knowledge-action feature value and a reference knowledge-action feature value as a knowledge-action influence factor;

[0033] determining a ratio of the language feature value and a reference language feature value as a language influence factor;

[0034] determining a weighted sum value of the knowledge-action influence factor and the language influence factor as a knowledge-action mapping feature value.

[0035] Further, the knowledge-action mapping module analyzes a knowledge-action mapping state, and determines a key teaching frame, wherein,

[0036] if the knowledge-action mapping feature value is greater than a knowledge-action mapping feature value threshold, the knowledge-action mapping state is an abnormal knowledge-action mapping tendency, and a frame corresponding to the knowledge-action mapping feature value is determined as a key teaching frame;

[0037] if the knowledge-action mapping feature value is less than or equal to the knowledge-action mapping feature value threshold, the knowledge-action mapping state is a normal knowledge-action mapping tendency, and the frame corresponding to the knowledge-action mapping feature value is determined as a normal teaching frame.

[0038] Further, the knowledge-action interaction module locates an abnormal feature, comprising,

[0039] normalizing the behavior time domain curve, the language time domain curve, and the reaction time domain curve;

[0040] placing the normalized behavior time domain curve, the language time domain curve, and the reaction time domain curve on the same coordinate axis, determining the horizontal coordinate of the coordinate axis as time and the vertical coordinate as relative intensity;

[0041] determining a behavior difference area of the behavior time domain curve and the reaction time domain curve;

[0042] determining a language difference area of the language time domain curve and the reaction time domain curve;

[0043] arranging the behavior difference area and the language difference area in descending order;

[0044] determining a feature corresponding to a difference area arranged in the first place as an abnormal feature.

[0045] Further, the knowledge-action interaction module calculates an abnormal feature coefficient, and determines a module to be called, wherein,

[0046] determining a difference value mean value of the difference area arranged in the first place and a reference difference area;

[0047] determining a ratio of the difference value mean value and the reference difference area as an abnormal feature coefficient;

[0048] if the abnormal feature coefficient is greater than an abnormal feature coefficient threshold, calling an associated regulation module;

[0049] If the abnormal feature coefficient is less than or equal to an abnormal feature coefficient threshold, a wisdom evaluation module is invoked.

[0050] Further, the correlation regulation module regulates the abnormal feature based on the abnormal feature coefficient, including,

[0051] selecting a corresponding modified behavior feature or modified language feature for the abnormal feature based on a behavior language database;

[0052] determining the modified behavior feature or the modified language feature as a modified feature;

[0053] The behavior language database includes behavior features and language features.

[0054] Compared with the prior art, the present application sets data storage modules, knowledge-action mapping modules, knowledge-action interaction modules, correlation regulation modules and wisdom evaluation modules, determines behavior effective areas and knowledge-action delay and knowledge-action matching degree of students, calculates knowledge-action feature values and language feature values for the behavior effective areas, to determine knowledge-action mapping feature values and key teaching frames, constructs a predetermined time period according to the key teaching frames, compares behavior time domain curves, language time domain curves and reaction time domain curves, locates abnormal features, calculates abnormal feature coefficients, determines modules to be invoked, the correlation regulation module generates modified features, and the wisdom evaluation module evaluates teaching effects. The present application determines knowledge-action mapping states through the language and behavior of teachers and the reaction data of students, to solve the problem that the existing evaluation methods are difficult to correlate teaching processes and effects, lack of pertinence, cause evaluation effects to deviate, and reduce analysis efficiency and evaluation accuracy.

[0055] Especially, by determining behavior effective areas, precise analysis of interaction feedback of target student groups and teachers is realized, thereby establishing a reliable data basis for evaluation of teaching effects. It can be understood that in actual teaching scenarios, most students are usually in a normal listening state, and only a few individuals have scattered attention. For a few individuals, teachers often remind them by directly calling their names or indicating specific areas. If undifferentiated analysis is still performed on all students at this time, not only will it cause redundant consumption of computing resources, but also will reduce system response and processing efficiency. Based on this, the present application determines behavior effective areas, focuses the analysis range on behavior performance of students in the area, to improve analysis accuracy and analysis efficiency, provides a data basis for subsequent teaching evaluation, and improves evaluation accuracy and evaluation efficiency.

[0056] Especially, by analyzing the knowledge-action delay and knowledge-action matching degree of the students in the behavior effective area, and combining the language data of the teachers to determine the knowledge-action mapping characteristic value, the objective and quantitative evaluation of the teacher-student interaction state is realized, which provides a theoretical basis and data basis for subsequent targeted further analysis. In actual conditions, when evaluating the teaching effect, the traditional method is often limited to one-sided analysis of the teacher behavior or the student state. This method is difficult to establish the real-time causal relationship between the teaching behavior and the student reaction, cannot capture the dynamic nature of the two-way feedback of classroom interaction, is easily affected by personal experience and accidental factors, and thus leads to significant deviation of the evaluation result. Based on this, the application considers to perform multi-modal data fusion analysis, collects and quantitatively analyzes the reaction data (including the knowledge-action delay and the matching degree) of the student end and the language data of the teacher end, calculates the knowledge-action mapping characteristic value which can comprehensively reflect the two-way interaction effect, and determines the key teaching frame, so as to provide a data basis for subsequent analysis of abnormal features affecting the key teaching frame, and improve the evaluation accuracy and evaluation efficiency.

[0057] Especially, by constructing the behavior time domain curve, the language time domain curve and the reaction time domain curve, and comparing and analyzing the three, the abnormal features in the teaching interaction process are accurately identified, the abnormal feature coefficient is calculated, and a data basis for subsequent generation of targeted correction features is provided. In actual conditions, the interaction abnormality may be caused by the low cognitive processing speed of the students, or by the deviation of the language expression or action demonstration of the teachers. It can be understood that the purpose of evaluating the teaching effect is to improve the classroom teaching quality, so after determining the interaction abnormality, the abnormal factors need to be further traced to make targeted correction. Based on this, the application considers the teacher and the student from two angles to analyze the reasons for the interaction abnormality, and quantifies the abnormal feature coefficient to dynamically call the subsequent processing module, thereby improving the evaluation accuracy and evaluation efficiency.

[0058] Especially, the teaching video segment corresponding to the predetermined time period is stored in the teaching effect evaluation database, which is used to evaluate the classroom teaching effect and to push knowledge to the students based on the teaching video segment. In actual conditions, when the students and the teachers have interaction problems, if the problem is caused by the students, it may be due to the problem understanding not being in place, the thinking speed being slow, etc. For this reason, if the method of adjusting the teaching behavior of the teachers is still used, the purpose of improving the teaching quality cannot be achieved. Based on this, the application considers to store the teaching video segment and accurately push the corresponding knowledge segment to the students according to their actual reactions in the classroom, so as to improve the teaching quality. BRIEF DESCRIPTION OF DRAWINGS

[0059] Figure 1 It is a structural schematic view of the classroom teaching effect evaluation system of the application embodiment.

[0060] Figure 2A diagram for determining a designated area of an embodiment of the invention;

[0061] Figure 3 A logic block diagram for determining a key teaching frame of an embodiment of the invention;

[0062] Figure 4 A logic block diagram for determining a module that needs to be called of an embodiment of the invention;

[0063] Wherein, student 1 (light-colored square), points to point 2, designated area 3 (black area). DETAILED DESCRIPTION

[0064] In order to make the objects and advantages of the present application clearer, the following further describes the present application with reference to embodiments; it should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0065] The preferred embodiments of the present application are described below with reference to the accompanying drawings. Those skilled in the art should understand that these embodiments are only used to explain the technical principles of the present application and are not intended to limit the protection scope of the present application.

[0066] It should be noted that, in the description of the present application, unless otherwise explicitly specified and limited, the term "connection" should be understood broadly, for example, it can be fixed connection, or detachable connection, or integrally connected; it can be mechanical connection, or electrical connection; it can be directly connected, or indirectly connected through an intermediate medium, or the internal communication of two elements. Those skilled in the art can understand the specific meaning of the above-mentioned term in the present application according to the specific circumstances.

[0067] Please refer to Figure 1 , Figure 1 A structure schematic diagram of a classroom teaching effect evaluation system of an embodiment of the invention. The classroom teaching effect evaluation system of the present application comprises:

[0068] A data storage module, which comprises a teacher end for obtaining behavior data and language data of a teacher to determine a behavior effective area, and a student end for obtaining reaction data of a student to determine a knowledge-action delay and a knowledge-action matching degree of the student;

[0069] A knowledge-action mapping module connected with the data storage module, for calculating a knowledge-action characteristic value based on the knowledge-action delay and the knowledge-action matching degree in the behavior effective area, calculating a language characteristic value based on the language data, determining a knowledge-action mapping characteristic value, analyzing a knowledge-action mapping state, and determining a key teaching frame;

[0070] The knowledge-action interaction module, which is connected to the knowledge-action mapping module, is used to extract behavioral data, language data and reaction data within a predetermined time period, compare the behavioral time-domain curve, language time-domain curve and reaction time-domain curve, locate abnormal features, calculate abnormal feature coefficients, and determine the modules to be called.

[0071] The associated control module, which is connected to the knowledge-action interaction module, is used to generate correction features based on the behavioral language database and send the correction features to the storage unit associated with the teacher's end.

[0072] The intelligent evaluation module, which is connected to the knowledge-action interaction module, is used to store the teaching video segments corresponding to the predetermined time period into the teaching effect evaluation database to evaluate the classroom teaching effect.

[0073] The behavioral data includes gesture amplitude and gesture trajectory, and the reaction data includes reaction time and line-of-sight target.

[0074] Specifically, there are no restrictions on the methods for acquiring behavioral and linguistic data. For example, video and audio of teachers can be collected by cameras deployed on the podium. Of course, those skilled in the art can also acquire the required data in other ways, as long as it is reasonable, which will not be elaborated here.

[0075] Specifically, there are no restrictions on the method of acquiring the reaction data. For example, the reaction data can be obtained by collecting videos of the student group through a wide-angle camera deployed at the back of the classroom. Of course, those skilled in the art can also obtain the required data in other ways, as long as it is reasonable. This will not be elaborated further.

[0076] Specifically, the length of the predetermined time period is not strictly limited. A preferred implementation is as follows: using the time corresponding to the key teaching frame as the base time point t0, a symmetrical time window is constructed. The defined interval of this time period is [t0-Δt, t0+Δt], where Δt is a preset time offset, which is set to 3 seconds in practice. Thus, the predetermined time period is an analysis window of 6 seconds, with a buffer of 3 seconds before and after, centered on the key teaching frame.

[0077] Specifically, the storage unit is used to store correction features so that teachers can make targeted adjustments to their abnormal behavior or language after class, which will not be elaborated further.

[0078] Specifically, the behavioral language database can be an open-source database containing standard teaching language and behavior, or it can be built based on historical data from several teachers within the school. It is understood that the historical data is all authorized, which will not be elaborated further here.

[0079] Specifically, the teaching effect evaluation database stores a plurality of teaching video segments, and the classroom teaching effect is evaluated by calling each teaching video segment.

[0080] Please refer to Figure 2 , Figure 2 The schematic diagram for determining the specified area of the embodiment of the application. Specifically, the data storage module determines the behavior effective area, including,

[0081] The pointing point of the behavior data is determined, and the minimum area containing the pointing point is determined as the specified area;

[0082] The gesture trajectory of the behavior data is identified, and the closed area surrounded by the gesture trajectory is determined;

[0083] The specified area and / or the closed area are determined as the behavior effective area.

[0084] It can be understood that the behavior effective area can be a pointing student area, or a pointing teaching tool area such as a blackboard. If the finger effective area points to the blackboard or other teaching tools, the students in the behavior effective area are all the students.

[0085] Specifically, the pointing point is the position pointed by the teacher's gesture. It can be understood that the pointing point can be one or more.

[0086] Specifically, by determining the behavior effective area, the accurate analysis of the interaction feedback between the target student group and the teacher is realized, thereby establishing a reliable data basis for the evaluation of the teaching effect. It can be understood that in the actual teaching scene, most students are usually in a normal listening state, and only a few individuals have their attention dispersed. For a few individuals, the teacher often reminds them by directly calling their names or indicating a specific area. If the whole class is still analyzed without discrimination at this time, it will not only lead to redundant consumption of computing resources, but also reduce the system response and processing efficiency. Based on this, the application determines the behavior effective area, focuses the analysis range on the behavior performance of the students in the area, to improve the analysis accuracy and efficiency, provides a data basis for subsequent teaching evaluation, and improves the evaluation accuracy and efficiency.

[0087] Specifically, the data storage module determines the knowledge-action delay and the knowledge-action matching degree of the students, including,

[0088] The absolute value of the difference between the reaction time and the end time of the language data is determined as the knowledge-action delay;

[0089] The ratio of the total number of the line of sight targets to the number of abnormal line of sight targets is determined as the knowledge-action matching degree;

[0090] The number of abnormal line-of-sight targets is different from the number of line-of-sight targets of the predetermined proportion.

[0091] It can be understood that the reaction moment is the moment when the student changes the action, for example, the student changes from lowering the head to lifting the head, and the lifting moment is the reaction moment.

[0092] Specifically, the specific value of the predetermined proportion is not limited, and in the implementation, the predetermined proportion is set to 30% of the total number of line-of-sight targets. Of course, the predetermined proportion can also be determined by the person skilled in the art, which will not be repeated here.

[0093] Specifically, the knowing and doing mapping module calculates the knowing and doing feature value, including,

[0094] The ratio of the knowing and doing delay to the reference knowing and doing delay is determined as a delay influence factor;

[0095] The ratio of the reference knowing and doing matching degree to the knowing and doing matching degree is determined as a matching influence factor;

[0096] The average of the sum of the inverse of the delay influence factor and the matching influence factor is determined as the knowing and doing feature value.

[0097] Specifically, the reference knowing and doing delay is calculated in advance, and the knowing and doing delays in a plurality of historical teaching processes are obtained in advance, and the average of each knowing and doing delay is determined as the reference knowing and doing delay.

[0098] Specifically, the reference knowing and doing matching degree is calculated in advance, and the knowing and doing matching degrees in a plurality of historical teaching processes are obtained in advance, and the average of each knowing and doing matching degree is determined as the reference knowing and doing matching degree.

[0099] Specifically, the knowing and doing mapping module calculates the language feature value, including,

[0100] The non-standard language in the language data is determined, and the ratio of the number of non-standard language to the total number of language is determined as the non-standard language ratio;

[0101] The logical fluency of the language data is determined.

[0102] The average of the sum of the inverse of the non-standard language ratio and the logical fluency is determined as the language feature value.

[0103] Specifically, the non-standard language may be a dialect, or it may be an incorrect language, of course, it may also be any vocabulary, pronunciation, grammar or expression manner that does not conform to the standard language, and the person skilled in the art can determine it by itself, which will not be repeated here.

[0104] Specifically, in the implementation, the language is regarded as a collection of several words or phrases, for example, there is an example sentence, "this example, its size is 3 cubic centimeters, this knowledge point is too important, everyone must master.", In the example sentence, there are 2 non-standard language, respectively, "3 cubic centimeters, too", all language has 12, respectively, "this, example, its size, is, 3 cubic centimeters, this, knowledge point, too, important, everyone, must, master", The non-standard language ratio in the example sentence is 16.7%, It can be understood that the division method of the example sentence is not limited, the person skilled in the art can divide according to the nature of the word or phrase, or directly according to the punctuation, which is reasonable, this will not be repeated.

[0105] Specifically, the calculation method of logical fluency is not limited, in the implementation, the logical fluency can be determined by calculating the semantic correlation of language data, the specific calculation steps are as follows,

[0106] The language data is cleaned and divided into basic units (such as words, phrases or sentences);

[0107] Calculate the association strength between each unit based on the semantic similarity model;

[0108] Determine the mean value of each association strength as the logical fluency.

[0109] Specifically, the know-how mapping module determines the know-how mapping feature value, including,

[0110] To determine the ratio of the know-how feature value and the reference know-how feature value as the know-how influence factor;

[0111] To determine the ratio of the language feature value and the reference language feature value as the language influence factor;

[0112] To determine the weighted sum value of the know-how influence factor and the language influence factor as the know-how mapping feature value.

[0113] Specifically, the reference know-how feature value is the know-how feature value corresponding to the reference know-how delay and the reference know-how matching degree.

[0114] Specifically, the reference language feature value is obtained by pre-computing, pre-acquire historical language feature values in several teaching processes, determine the mean value of each historical language feature value as the reference language feature value.

[0115] Specifically, the sum of the weight coefficients of the know-how influence factor and the language influence factor is 1, when configuring the weight coefficient, considering that the interaction state of the teacher and the student is affected by both parties, so the weight coefficient of the know-how influence factor is set to 0.5, and the weight coefficient of the language influence factor is set to 0.5.

[0116] Specifically, by analyzing the knowledge-action delay and knowledge-action matching degree of students in the behavior effective area, and combining the language data of teachers to determine the knowledge-action mapping feature value, the objective and quantitative evaluation of the teacher-student interaction state is realized, which provides a theoretical basis and data basis for subsequent targeted further analysis. In actual situations, when evaluating the teaching effect, the traditional method is often limited to one-sided analysis of teacher behavior or student state. This method is difficult to establish a real-time causal relationship between teaching behavior and student response, cannot capture the dynamic nature of two-way feedback of classroom interaction, is easily affected by personal experience and accidental factors, and thus leads to significant deviation in the evaluation result. Based on this, the present application considers multi-modal data fusion analysis, collects and quantitatively analyzes the reaction data (including knowledge-action delay and matching degree) of the student end and the language data of the teacher end, calculates the knowledge-action mapping feature value which can comprehensively reflect the two-way interaction effect, and determines the key teaching frame, thereby providing a data basis for subsequent analysis of abnormal features affecting the key teaching frame, and improving the evaluation accuracy and evaluation efficiency.

[0117] Please refer to Figure 3 , Figure 3 The logic block diagram for determining the key teaching frame of the present application embodiment. Specifically, the knowledge-action mapping module analyzes the knowledge-action mapping state and determines the key teaching frame, wherein,

[0118] If the knowledge-action mapping feature value is greater than the knowledge-action mapping feature value threshold, the knowledge-action mapping state is an abnormal knowledge-action mapping tendency, and the frame corresponding to the knowledge-action mapping feature value is determined as the key teaching frame.

[0119] If the knowledge-action mapping feature value is less than or equal to the knowledge-action mapping feature value threshold, the knowledge-action mapping state is a normal knowledge-action mapping tendency, and the frame corresponding to the knowledge-action mapping feature value is determined as the ordinary teaching frame.

[0120] Specifically, the knowledge-action mapping feature value threshold represents a boundary of the teacher-student interaction state, which is obtained by pre-computing. The knowledge-action mapping feature values that exist abnormally in several teaching processes are pre-computed, and the product of the mean value and the precision coefficient of each knowledge-action mapping feature value is determined as the knowledge-action mapping feature value threshold. The precision coefficient is selected within the interval [0.8, 1]. In actual situations, in order to improve the accuracy and precision of the analysis, the precision coefficient is determined as 0.9.

[0121] Specifically, the knowledge-action interaction module locates abnormal features, including,

[0122] to normalize the behavior time domain curve, the language time domain curve and the reaction time domain curve;

[0123] placing the normalized behavior time-domain curve, the language time-domain curve and the reaction time-domain curve on the same coordinate axis, determining the horizontal coordinate of the coordinate axis as time and the vertical coordinate as relative intensity;

[0124] determining the behavior difference area of the behavior time-domain curve and the reaction time-domain curve;

[0125] determining the language difference area of the language time-domain curve and the reaction time-domain curve;

[0126] ranking the behavior difference area and the language difference area in descending order;

[0127] determining the feature corresponding to the difference area ranked first as an abnormal feature.

[0128] Specifically, the construction method of the behavior time-domain curve, the language time-domain curve and the reaction time-domain curve and the steps of normalization are as follows,

[0129] constructing the initial behavior time-domain curve based on the amplitude of the gesture and time;

[0130] constructing the initial language time-domain curve based on the keyword frequency and time;

[0131] constructing the initial reaction time-domain curve based on the head-up rate of the student and time;

[0132] performing minimum-maximum normalization on the intensity value of each curve on the time axis, and mapping all values to the interval [0, 1];

[0133] completing the construction and normalization of the behavior time-domain curve, the language time-domain curve and the reaction time-domain curve.

[0134] Specifically, the difference area is the absolute value of the difference between the vertical coordinate values of the two boundary time points of the predetermined time period between the two time-domain curves, integrated along the time axis

[0135] Please refer to Figure 4 , Figure 4 The logic block diagram of the modules required to be called by the embodiment of the application is shown in FIG. 6. Specifically, the know-do interaction module calculates the abnormal feature coefficient and determines the modules required to be called, wherein,

[0136] determining the average value of the difference between the difference area ranked first and the reference difference area;

[0137] determining the ratio of the average value of the difference and the reference difference area as the abnormal feature coefficient;

[0138] if the abnormal feature coefficient is greater than the abnormal feature coefficient threshold, calling the correlation regulation module;

[0139] If the abnormal feature coefficient is less than or equal to an abnormal feature coefficient threshold, a wisdom evaluation module is invoked.

[0140] Specifically, the reference difference area is pre-calculated, and a plurality of difference areas in a normal teaching process are pre-acquired, and the mean of each difference area is determined as the reference difference area.

[0141] Specifically, the abnormal feature coefficient threshold represents a boundary for generating a correction feature, and is pre-calculated, and a plurality of abnormal feature coefficients requiring the generation of a correction feature are pre-acquired, and the product of the mean of each abnormal feature coefficient and an abnormal accuracy coefficient is determined as the abnormal feature coefficient threshold, and the abnormal accuracy coefficient is selected in the interval [0.8, 1.0], and in actual situations, to improve the analysis accuracy, the abnormal accuracy coefficient is determined as 0.9.

[0142] Specifically, by constructing a behavior time domain curve, a language time domain curve and a reaction time domain curve, and comparing and analyzing the three, the abnormal feature in the teaching interaction process is accurately identified, the abnormal feature coefficient is calculated, and a data basis is provided for subsequent generation of a targeted correction feature. In actual situations, the interaction abnormality may be caused by a low cognitive processing speed of the student, or by a deviation in the language expression or action demonstration of the teacher. It can be understood that the purpose of evaluating the teaching effect is to improve the classroom teaching quality, so after determining the interaction abnormality, the abnormal factor needs to be further traced to make targeted corrections. Based on this, the present application considers both the teacher and the student, analyzes the reasons for the interaction abnormality, and quantifies the abnormal feature coefficient to dynamically invoke the subsequent processing module, thereby improving the evaluation accuracy and efficiency.

[0143] Specifically, the correlation regulation module regulates the abnormal feature based on the abnormal feature coefficient, including,

[0144] to select a corresponding correction behavior feature or a correction language feature for the abnormal feature based on a behavior language database;

[0145] determining the correction behavior feature or the correction language feature as a correction feature;

[0146] The behavior language database includes behavior features and language features.

[0147] It can be understood that if the abnormal feature is a behavior feature corresponding to behavior data, the correction behavior feature corresponding to the behavior feature exists in the behavior language database, and the correction behavior feature is determined as the correction feature.

[0148] Specifically, the application stores the teaching video segment corresponding to the predetermined time period into a teaching effect evaluation database, to evaluate the classroom teaching effect, to push knowledge to students based on the teaching video segment, and in actual situations, when the student and the teacher appear interactive problems, if the factor causing the problem is the student, it may be because the problem is not fully understood, the thinking speed is slow and other factors, for this reason, if the method of adjusting the teacher's teaching behavior is still used, it will not achieve the purpose of improving the teaching quality, based on this, the application considers storing the teaching video segment, and according to the actual reaction of the student in the classroom, the corresponding knowledge segment is accurately pushed to the student, to improve the teaching quality.

[0149] So far, the technical solutions of the application have been described in combination with the preferred embodiments shown in the drawings, but those skilled in the art can easily understand that the protection scope of the application is obviously not limited to these specific embodiments. Those skilled in the art can make equivalent changes or replacements to the related technical features without departing from the principles of the application, and the technical solutions after the changes or replacements will fall within the protection scope of the application.

[0150] The above description is only the preferred embodiments of the application and is not used to limit the application; for those skilled in the art, the application can have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the application shall be included in the protection scope of the application.

Claims

1. A classroom teaching effect evaluation system characterized by, The method comprises the following steps: A data storage module comprises a teacher terminal for obtaining behavior data and language data of a teacher to determine a behavior effective area, and a student terminal for obtaining reaction data of a student to determine a behavior-knowledge delay and a behavior-knowledge matching degree of the student; A behavior-knowledge mapping module connected with the data storage module is used to calculate a behavior-knowledge feature value based on the behavior-knowledge delay and the behavior-knowledge matching degree within the behavior effective area, calculate a language feature value based on the language data, determine a behavior-knowledge mapping feature value, analyze a behavior-knowledge mapping state, and determine a key teaching frame; A behavior-knowledge interaction module connected with the behavior-knowledge mapping module is used to extract behavior data, language data and reaction data within a predetermined time period, compare behavior time domain curves, language time domain curves and reaction time domain curves, locate abnormal features, calculate abnormal feature coefficients, and determine a module to be called; An association regulation module connected with the behavior-knowledge interaction module is used to generate a correction feature based on a behavior language database, and send the correction feature to a storage unit associated with the teacher terminal; A smart evaluation module connected with the behavior-knowledge interaction module is used to store a teaching video segment corresponding to the predetermined time period to a teaching effect evaluation database, and evaluate a classroom teaching effect; The behavior data comprises gesture amplitude and gesture trajectory, and the reaction data comprises reaction time and a line of sight target; The data storage module determines the behavior-knowledge delay and the behavior-knowledge matching degree of the student, comprising: The absolute value of the difference between the reaction time and the end time of the language data is determined as the behavior-knowledge delay; The ratio of the total number of the line of sight targets to the number of abnormal line of sight targets is determined as the behavior-knowledge matching degree; The number of abnormal line of sight targets is different from the number of line of sight targets with a predetermined proportion; The behavior-knowledge interaction module calculates the abnormal feature coefficients and determines the module to be called, wherein: The average value of the difference between the first difference area and the reference difference area is determined; The ratio of the average value to the reference difference area is determined as the abnormal feature coefficient; If the abnormal feature coefficient is greater than an abnormal feature coefficient threshold, the association regulation module is called; If the abnormal feature coefficient is less than or equal to the abnormal feature coefficient threshold, the smart evaluation module is called; The association regulation module regulates the abnormal features based on the abnormal feature coefficients, comprising: Based on the behavior language database, the abnormal features are selected to have corresponding correction behavior features or correction language features; The correction behavior features or the correction language features are determined as correction features. The behavior language database comprises behavior features and language features.

2. The classroom instruction effectiveness evaluation system according to claim 1, characterized by, The data storage module determines the behavior effective area, comprising: The pointing point of the behavior data is determined, and the smallest area containing the pointing point is determined as a specified area; The gesture trajectory of the behavior data is identified, and a closed area surrounded by the gesture trajectory is determined; The specified area and / or the closed area are determined as the behavior effective area.

3. The classroom instruction effectiveness evaluation system according to claim 1, characterized by The behavior-knowledge mapping module calculates the behavior-knowledge feature value, comprising: The ratio of the behavior-knowledge delay to a reference behavior-knowledge delay is determined as a delay influence factor. ​ determining a ratio of the reference knowledge-action matching degree and the knowledge-action matching degree as a matching influence factor; determining an average of a sum of the delay influence factor and an inverse of the matching influence factor as a knowledge-action characteristic value.

4. The classroom instruction effectiveness evaluation system according to claim 1, characterized by The knowledge-action mapping module calculates a language characteristic value, including, determining non-standard expressions in the language data, determining a ratio of a number of the non-standard expressions and a total number of expressions as a non-standard expression ratio; determining a logical fluency of the language data; determining an average of a sum of the non-standard expression ratio and an inverse of the logical fluency as the language characteristic value.

5. The classroom instruction effectiveness evaluation system according to claim 1, characterized by, The knowledge-action mapping module determines a knowledge-action mapping characteristic value, including, determining a ratio of the knowledge-action characteristic value and a reference knowledge-action characteristic value as a knowledge-action influence factor; determining a ratio of the language characteristic value and a reference language characteristic value as a language influence factor; determining a weighted sum of the knowledge-action influence factor and the language influence factor as the knowledge-action mapping characteristic value.

6. The classroom instruction effectiveness evaluation system according to claim 1, characterized by The knowledge-action mapping module analyzes a knowledge-action mapping state, and determines a key teaching frame, wherein, if the knowledge-action mapping characteristic value is greater than a knowledge-action mapping characteristic value threshold, the knowledge-action mapping state is an abnormal knowledge-action mapping tendency, and a frame corresponding to the knowledge-action mapping characteristic value is determined as the key teaching frame; if the knowledge-action mapping characteristic value is less than or equal to the knowledge-action mapping characteristic value threshold, the knowledge-action mapping state is a normal knowledge-action mapping tendency, and the frame corresponding to the knowledge-action mapping characteristic value is determined as a common teaching frame.

7. The classroom instruction effectiveness evaluation system according to claim 1, characterized by The knowledge-action interaction module locates an abnormal feature, including, normalizing the behavior time domain curve, the language time domain curve and the reaction time domain curve; placing the normalized behavior time domain curve, the language time domain curve and the reaction time domain curve on a same coordinate axis, determining a horizontal coordinate of the coordinate axis as time and a vertical coordinate of the coordinate axis as relative intensity; determining a behavior difference area of the behavior time domain curve and the reaction time domain curve; determining a language difference area of the language time domain curve and the reaction time domain curve; arranging the behavior difference area and the language difference area in descending order; determining a feature corresponding to a difference area arranged in the first place as the abnormal feature.

Images