Mathematics learning system and method through deep analysis of mathematical problem-solving data

Abstract

The present invention relates to a mathematics learning system and method through in-depth analysis of mathematics problem-solving data. More specifically, it relates to a mathematics learning system and method through in-depth analysis of mathematics problem-solving data that analyzes a learner's mathematics problem-solving data to derive the learner's strengths and weaknesses by comparing and evaluating logical development methods, concept application methods, calculation accuracy, and solution time with model solutions, and automatically provides customized learning content for each learner based on the derived evaluation results.

Description

Technology Field

[0001] The present invention relates to a mathematics learning system and method through in-depth analysis of mathematics problem-solving data. More specifically, it relates to a mathematics learning system and method through in-depth analysis of mathematics problem-solving data that analyzes a learner's mathematics problem-solving data to derive the learner's strengths and weaknesses by comparing and evaluating logical development methods, concept application methods, calculation accuracy, and solution time with model solutions, and automatically provides customized learning content for each learner based on the derived evaluation results. Background Technology

[0003] Generally, among the various subjects students learn in school, mathematics is the one they find most difficult to study.

[0004] Mathematics fundamentally deals with many abstract concepts—such as numbers, symbols, and formulas—that are not directly related to concrete objects or situations in the real world. Since it is a discipline where basic concepts build upon each other to advance to advanced concepts, if students move on to the next stage without sufficiently understanding the basic concepts, subsequent learning becomes increasingly difficult.

[0005] Furthermore, recent math question types are increasingly requiring problem-solving skills rather than simply relying on formulas or calculations; students may face significant difficulties in solving these problems if they lack experience in analyzing problems, finding solutions, and applying those methods.

[0006] When students repeatedly struggle to solve math problems, their confidence drops and they gradually lose interest in learning. Due to this problem, the number of students who have given up on math, known as "math dropouts," has been increasing recently.

[0007] Recently, with the advancement of modern educational technology, there has been increasing interest in personalized learning systems tailored to the characteristics of individual learners. In particular, for subjects requiring theoretical and logical thinking, such as mathematics, the importance of learning methods that consider the learner's individual level and learning patterns is being further emphasized.

[0008] Traditional mathematics education, based on a uniform curriculum and teaching methods, has struggled to adequately reflect individual differences among learners. This leads to difficulties in solving mathematical problems and a decline in learning motivation, which can result in a negative perception of mathematics.

[0009] Existing mathematics learning systems have primarily operated by relying on methods of solving fixed problems or by providing feedback based on results after evaluation.

[0010] However, these systems not only have limitations in analyzing learners' individual strengths and weaknesses in real time and reflecting them in the learning process, but also struggle to provide personalized feedback tailored to the learner's abilities or level of understanding, often failing to support continuous learning motivation and effective learning improvement.

[0011] In addition, there is a lack of systems or methods to quantitatively evaluate the logical development, concept application methods, and accuracy and efficiency of calculation processes in the process of solving problems by learners. Therefore, there are limitations in that it is difficult to individually analyze learners even if they repeatedly make mistakes in specific types of problems, and it is also difficult to provide customized learning that can compensate for individual weaknesses. Prior art literature

[0013] 1. Korean Patent Publication No. 10-2636079 (Registered on Feb. 6, 2024) The problem to be solved

[0014] The present invention has been devised to solve the problems of the aforementioned prior art. The objective of the present invention is to provide a mathematics learning system and method through in-depth analysis of mathematics problem-solving data, which analyzes a learner's mathematics problem-solving data to compare and evaluate the learner's logical development method, concept application method, calculation accuracy, and solution time with a model solution to derive the learner's strengths and weaknesses, and automatically provides customized learning content for each learner based on the derived evaluation results. means of solving the problem

[0016] The present invention for achieving the above objectives is,

[0017] A mathematics learning system may include: a database that stores mathematics problems to be provided to learners, classified by unit, type, and difficulty level; a data collection unit that collects video data captured by the first and second cameras, including a first camera that captures the learner's problem-solving process from the front and a second camera that captures the learner's problem-solving process from the top; and a main server that is connected to the database and the data collection unit via a network, analyzes the video data collected through the data collection unit, and provides customized feedback for each learner.

[0018] At this time, the main server may include: a data analysis unit that analyzes video data collected from the data collection unit; a comparison and evaluation unit that derives the learner's strengths and weaknesses in the problem-solving process by comparing the analysis results from the data analysis unit with best practices; and a feedback providing unit that provides feedback to compensate for weaknesses by considering the learner's strengths and weaknesses derived from the comparison and evaluation unit.

[0019] Additionally, the data analysis unit may include: an eye gaze analysis module that analyzes the learner's eye movement during the problem-solving process using video data captured by the first camera; a problem-solving process analysis module that analyzes the problem-solving process by detecting the position of handwriting during the problem-solving process using video data captured by the second camera; and a first visualization module that visualizes the analysis results by the eye gaze analysis module and the problem-solving process analysis module.

[0020] In addition, the comparison and evaluation unit may include: a solution strategy comparison module that compares the learner's problem-solving strategy with a best practice using the analysis results from the data analysis unit; a calculation process comparison module that compares the learner's calculation process with a best practice using the analysis results from the data analysis unit; a solution time comparison module that compares the learner's problem-solving time with a best practice using the analysis results from the data analysis unit; and a second visualization module that visualizes and displays the comparison results from the solution strategy comparison module, the calculation process comparison module, and the solution time comparison module.

[0021] Additionally, the feedback providing unit may include: a problem type determination module that selects and provides a type of mathematical problem from a database that can compensate for the learner's weaknesses derived from the comparison and evaluation unit; a numerical transformation module that utilizes AI functions to transform numerical information included in the mathematical problem provided by the problem type determination module; and a similar problem generation module that generates a similar problem by applying the numerical value transformed by the numerical transformation module to the mathematical problem selected by the problem type determination module.

[0022] Meanwhile, the mathematics learning method using learner analysis data according to the present invention is,

[0023] The method may include: a data collection step of acquiring video data capturing the learner's problem-solving process using a first camera installed in front of the learner and a second camera installed above the learner; a data analysis step of analyzing the learner's problem-solving process through a data analysis unit of a main server using the video data collected in the data collection step; a comparison and evaluation step of deriving the learner's strengths and weaknesses in the problem-solving process by comparing the analysis results from the data analysis step with a model case in which a model problem-solving process was filmed; and a feedback provision step of providing feedback to supplement weaknesses by considering the learner's strengths and weaknesses derived in the comparison and evaluation step. Effects of the invention

[0025] According to the present invention, by analyzing a learner's mathematical problem-solving data and comparing and evaluating the logical development method, concept application method, calculation accuracy, and solution time with a model solution, the learner's strengths and weaknesses are derived, and based on the derived evaluation results, customized learning content for each learner can be automatically provided, which has an excellent effect.

[0026] In addition, according to the present invention, learners can identify shortcomings in the problem-solving process and intuitively check situations that are being improved through feedback, thereby boosting learners' motivation and stimulating their interest, while simultaneously improving learning efficiency and learning productivity. Brief explanation of the drawing

[0028] FIG. 1 is a diagram schematically showing the configuration of a mathematics learning system through deep analysis of mathematics problem-solving data according to the present invention. FIG. 2 is a conceptual drawing of the present invention shown in FIG. 1. FIGS. 3(a) and (b) are diagrams exemplarily illustrating a mathematical problem selected by the problem type determination module of the present invention shown in FIG. 2 and a similar problem generated by the similar problem generation module. FIG. 4 is a flowchart sequentially illustrating a mathematical learning method through in-depth analysis of mathematical problem-solving data according to the present invention. Specific details for implementing the invention

[0029] The embodiments of the present disclosure are illustrative for the purpose of explaining the technical concept of the present disclosure. The scope of rights according to the present disclosure is not limited to the embodiments presented below or the specific description thereof.

[0030] All technical and scientific terms used in this disclosure, unless otherwise defined, have the meaning generally understood by those skilled in the art to which this disclosure pertains. All terms used in this disclosure are selected for the purpose of further clarifying this disclosure and are not selected to limit the scope of the rights under this disclosure.

[0031] Expressions such as “comprising,” “comprising,” “having,” etc. used in this disclosure should be understood as open-ended terms implying the possibility of including other embodiments, unless otherwise stated in the phrase or sentence containing such expressions.

[0032] Unless otherwise stated, singular expressions described in this disclosure may include a plural meaning, and this applies likewise to singular expressions described in the claims.

[0034] Hereinafter, preferred embodiments of a mathematics learning system and method through deep analysis of mathematics problem-solving data according to the present invention will be described in detail with reference to the attached drawings.

[0035] FIG. 1 is a schematic diagram showing the configuration of a mathematics learning system through deep analysis of mathematics problem-solving data according to the present invention, FIG. 2 is a conceptual diagram showing the present invention shown in FIG. 1, FIG. 3 (a) and (b) are exemplary diagrams showing a mathematics problem selected by the problem type determination module and a similar problem generated by the similar problem generation module among the present invention shown in FIG. 2, FIG. 4 is a flowchart showing a mathematics learning method through deep analysis of mathematics problem-solving data according to the present invention in sequence.

[0037] The present invention relates to a mathematics learning system and method through deep analysis of mathematics problem-solving data, which analyzes a learner's mathematics problem-solving data to derive the learner's strengths and weaknesses by comparing and evaluating the logical development method, concept application method, calculation accuracy, and solution time with a model solution, and automatically provides customized learning content for each learner based on the derived evaluation results. First, the mathematics learning system (10) according to the present invention may include, as shown in FIG. 1, a database (100), a data collection unit (200), a main server (300), and a learner terminal (400).

[0038] First, the above database (100) serves to integrally store and manage information regarding mathematical problems and related content provided in the present invention, and the mathematical problems stored in the above database (100) can be provided to a learner terminal (400) through the main server (300) to be described later.

[0039] That is, the information that can be stored in the above database (100) may include math problems of various types and difficulty levels, best practice information, learner information, etc., and each of the information may be stored separately in a separate DB.

[0040] To explain in more detail, the above database (100) may include a problem DB (110), a best practice DB (120), a learner DB (130), etc., as shown in FIG. 2. First, the problem DB (110) is configured to store math problems that may be included in a problem sheet to be transmitted to a learner terminal (400) by the main server (300) to be described later. Various existing types of math problems, as well as problems newly generated by the similar problem generation module (336) of the main server (300) to be described later, can be stored separately by unit, type, and difficulty level.

[0041] Next, the above-mentioned best practice DB (120) is configured to store problem-solving video data of high-achieving students or teachers captured by the first and second cameras (210, 220) to be described later, and a more detailed explanation thereof will be provided later.

[0042] Next, the above-mentioned learner DB (130) is configured to store information of learners using the learner terminal (400) described later. Student information entered into the main server (300) through the learner terminal (400) can be transmitted to the database (100) via a network and stored in the learner DB (130).

[0043] In addition, the above learner DB (130) may store the strengths and weaknesses of learners evaluated by the comparison and evaluation unit (320) described later, or the causes of inefficient problem-solving, and may additionally store feedback content provided by the feedback provision unit (330) and improvements based thereon.

[0044] Next, the data collection unit (200) is configured to collect information including phenomena occurring during the process of learners solving math problems, namely problem-solving habits, solution strategies, solution processes, and time taken, and may include first and second cameras (210, 220) that film the process of learners solving problems.

[0045] To explain in more detail, the first and second cameras (210, 220) are respectively installed in front of and above the learner solving the problem, so that the learner's problem-solving appearance can be captured from different directions. The video captured by the first camera (210) can be used to analyze the learner's eye movement in the data analysis unit (310) of the main server (300) to be described later, and the video captured by the second camera (220) can be utilized for the analysis of the solution stage in the data analysis unit (310).

[0046] That is, as will be described later, when a learner solves a problem sheet such as a test paper (hereinafter collectively referred to as 'problem sheet'), the learner only looks at the problem sheet and their own solution. Therefore, the data collection unit (200) can use multiple cameras (210, 220) to photograph the learner's problem-solving process from various directions, and then transmit the captured video to the main server (300). Then, the data analysis unit (310) of the main server (300) can analyze the learner's eye movements and handwriting positions to automatically distinguish the detailed steps of the solution process and check the time required for each step.

[0047] Next, the main server (300) is connected to the database (100) and the data collection unit (200) via a network and analyzes video data collected through the data collection unit (200) to provide customized feedback for each learner, including customized math problems stored in the database (100). As shown in FIG. 2, the configuration may include a data analysis unit (310), a comparison and evaluation unit (320), and a feedback provision unit (330).

[0048] First, the data analysis unit (310) is configured to analyze image data collected from the data collection unit (200) and may include a gaze analysis module (312), a solution process analysis module (314), and a first visualization module (316).

[0049] To explain in more detail, the eye-tracking analysis module (312) is configured to analyze the movement of the learner's eyes during the problem-solving process using video data captured by the first camera (210), and can measure the time the learner's eyes stay on a specific part by analyzing the movement of the learner's pupils in the video.

[0050] That is, the gaze analysis module (312) can identify the learner's pupil in the video data captured by the first camera (210), and then analyze the movement of the pupil to estimate the direction in which the learner's gaze is directed, i.e., the gaze vector.

[0051] Through this, by analyzing the learner's eye angle and eye movement patterns, it is possible to confirm where the gaze rested on specific parts of the problem sheet, such as problem conditions, numbers, or graphs; and if pupil movement is detected, the duration of the gaze resting on a specific part can be determined by calculating the frame time difference between the movements.

[0052] In addition, the above-mentioned gaze analysis module (312) can perform learning to estimate the pupil and gaze direction in an image using a known CNN (Convolutional Neural Network) model, thereby improving the accuracy of the learner's gaze analysis.

[0053] Next, the above-mentioned problem-solving process analysis module (314) is configured to analyze the problem-solving process by detecting the handwriting position in the problem-solving process using video data captured by the second camera (220), and can distinguish the learner's problem-solving process in stages by combining it with the learner's pupil movement analyzed by the gaze analysis module (312).

[0054] For example, the above-mentioned solution process analysis module (314) can divide the learner's problem-solving process into four stages: problem reading stage, concept organization stage, operation stage, and organization review stage. When the learner's gaze confirmed from the image of the first camera (210) is directed toward the problem text, and the handwriting position confirmed from the image of the second camera (220) is the problem text or the learner is not writing, it can be determined as the problem reading stage.

[0055] Next, if the learner's gaze moves away from the problem text, the position of the notes changes, or the learner is not taking notes, it can be determined as the concept organization stage, where the learner recalls related concepts from the problem and organizes formulas; subsequently, if the learner begins taking notes in earnest—that is, solving the problem—and their gaze moves between the problem text and the notes section, it can be determined as the operation stage.

[0056] In addition, if the learner's gaze returns to the text of the question and the handwritten notes also move toward the text of the question, it can be determined as the review stage where the final answer is decided and marked.

[0057] The above solution process analysis module (314) can calculate the time required for each step by analyzing the frames of the video data captured by the first and second cameras (210, 220) when the above four-step solution process analysis is completed.

[0058] Next, the first visualization module (316) is configured to visualize the analysis results of the gaze analysis module (312) and the solution process analysis module (314), and as shown in (Table 1) below, the time required for each solution step can be displayed in the form of a table or a graph.

[0059]

[0060] In addition, the first visualization module (316) can visualize the area where the learner’s gaze rested and the handwritten part during problem solving in the form of a heatmap, and can be configured to transmit the heatmap to the learner terminal (400) so that learners can intuitively check the level of concentration during the problem solving process.

[0061] Additionally, although not illustrated, the mathematics learning system (10) according to the present invention may include a monitoring means for displaying the contents visualized by the first visualization module (316) and the second visualization module (328) to be described later.

[0062] Next, the comparison and evaluation unit (320) is configured to derive the strengths and weaknesses of the learner by comparing the analysis results from the data analysis unit (310) with best practices, and the best practices may include the analysis results of solutions from high-achieving students or the best practices of teachers.

[0063] That is, the problem-solving process of high-achieving students or teachers is captured using the first and second cameras (210, 220) of the data collection unit (200), and the captured video data is stored separately as a model case in the database (100). By comparing the analysis results of the learner analyzed by the data analysis unit (310) with the model case stored in the database (100), the strengths and weaknesses of the learner can be derived.

[0064] At this time, key factors for evaluating the learner's analysis results may include problem-solving strategies, calculation processes, and time efficiency.

[0065] To explain in more detail, the comparison and evaluation unit (320) may include a solution strategy comparison module (322), an operation process comparison module (324), a solution time comparison module (326), and a second visualization module (328). The solution strategy comparison module (322) is configured to compare the learner's problem-solving strategy with a model case using the analysis results from the data analysis unit (310), and can compare how the same math problem was solved.

[0066] For example, the solution strategy comparison module (322) analyzes which method is more efficient for solving a mathematical problem that can be solved algebraically or geometrically, and then checks whether the solution method of the learner confirmed from the video data captured by the second camera (220) in the solution process analysis module (314) is the same as the solution method in the model case.

[0067] At this time, the solution strategy comparison module (322) can present a faster and more accurate solution method by utilizing AI functions such as pattern recognition to analyze algebraic or geometric approaches according to the characteristics of the problem and comparing efficiency such as computational complexity.

[0068] Next, the above-mentioned calculation process comparison module (324) is configured to compare the learner's calculation process with a model case using the analysis results from the data analysis unit (310), thereby allowing verification of the complexity of the calculation in the learner's problem-solving process, that is, whether unnecessary calculations are included.

[0069] The above calculation process comparison module (324) can evaluate the calculation complexity of a learner by comparing the calculation process for solving a problem, which is confirmed from the image data captured by the second camera (220) in the above solution process analysis module (314), with the calculation process of the same problem that can be confirmed from a model case, and the number of calculations can be used for such evaluation of calculation complexity.

[0070] That is, the above calculation process comparison module (324) compares the number of calculations of the learner for problem solving with the number of calculations in the model case, and if the number of calculations of the learner is greater, it can identify unnecessary calculations in the learner's solution process through comparison with the model case and analyze the reason why unnecessary calculations are included.

[0071] To explain in more detail, the calculation process comparison module (324) can detect duplicate calculations or unnecessary variables in the learner's calculation process, reduce the number of calculations through the removal of identities or duplicate constants and factorization, and improve the accuracy of recognizing unnecessary calculations in the learner's calculation process through machine learning.

[0072] Next, the above-mentioned solution time comparison module (326) is configured to compare the time taken for a learner to solve a problem with a model case using the analysis results from the data analysis unit (310). By measuring the time taken for each solution step from the problem-solving video data of the model case stored in the database (100) and comparing it with the time taken for each solution step of the learner calculated by the solution process analysis module (314), inefficient solution steps can be identified.

[0073] At this time, if there is a problem-solving step of a learner that takes excessive time compared to a best practice case, the solution time comparison module (326) can utilize an AI function to analyze in detail the video data of the corresponding section, that is, the video data of the learner captured by the first and second cameras (210, 220) and the video data of the best practice case, to identify the cause of the longer time taken, such as insufficient understanding of the problem, insufficient solution strategy, or inefficient calculation.

[0074] In addition, the solution time comparison module (326) can identify the cause of repeated occurrences of solution steps that take a long time compared to the best practice through video data analysis, such as inefficient eye movement or motion patterns occurring repeatedly or inefficient calculation methods being used repeatedly, and can store the identified causes in the learner DB (130) provided in the database (100).

[0075] Next, the second visualization module (328) is configured to visualize and display the comparison results from the solution strategy comparison module (322), the calculation process comparison module (324), and the solution time comparison module (326), and can visualize the learner's shortcomings or inefficiencies in the form of a table or graph by comparing them with a best practice or optimal solution path.

[0076] For example, the second visualization module (328) above can compare the learner's solution and the model solution item by item and analyze the differences and display them in a table form, as shown in (Table 2) below.

[0077]

[0079] In addition, the second visualization module (328) can provide video data of best practices for each question, thereby enabling learners to intuitively realize their shortcomings, and such video data of best practices or comparison results with best practices can be transmitted to a learner terminal (400) via a network.

[0080] Next, the feedback providing unit (330) is configured to provide feedback to compensate for weaknesses by considering the strengths and weaknesses of the learner derived from the comparison and evaluation unit (320), and can provide learner-specific feedback that can improve inefficient parts of the learner's solution process derived through the comparison and evaluation unit (320) by utilizing AI functions.

[0081] For example, the feedback providing unit (330) can learn the learner's inefficient problem-solving patterns and continuously provide feedback that can improve the learner's problem-solving method through machine learning such as reinforcement learning.

[0082] In addition, the feedback providing unit (330) can provide individual customized training content by creating and providing additional learning problems to the learner to compensate for the learner's weaknesses.

[0083] To explain in more detail, the feedback providing unit (330) may include a problem type determination module (332), a numerical transformation module (334), and a similar problem generation module (336). First, the problem type determination module (332) may select and retrieve math problems from the problem DB (110) of the database (100) to compensate for the learner's weaknesses.

[0084] That is, when a learner's weakness is identified through comparison with a best practice in the comparison and evaluation unit (320), the problem type determination module (332) can select and retrieve a type of math problem from the problem DB (110) of the database (100) that can compensate for the identified weakness of the learner.

[0085] Next, the numerical transformation module (334) is configured to generate similar problems that learners can train to compensate for weaknesses by transforming the numerical values ​​included in the math problem selected by the problem type determination module (332), and can randomly transform the numerical information included in the problem conditions and question content by utilizing AI functions.

[0086] For example, if the math problem selected by the problem type determination module (332) is (a) of FIG. 3, the numerical transformation module (334) can randomly transform the value of b4 = 6 included in the problem condition into c3 = 4, and transform b6 = 3b5 - 24 into a form such as c5 = 2c4 + 10, and at the same time change the value of b7 included in the question content to the value of c6, randomly generate options corresponding to incorrect answers based on the calculation result of c6.

[0087] At this time, the numerical transformation module (334) can maintain a constant problem difficulty by setting the transformed coefficient to an integer when the coefficient included in the problem condition is an integer, and can be configured so that the similar problem generation module (336) described later can form multiple similar problems for a single original problem with different difficulty levels by using methods such as transforming the coefficient into a rational number to adjust the difficulty.

[0088] Next, the above-mentioned similar problem generation module (336) is configured to generate similar problems by applying numerical values ​​modified by the numerical transformation module (334) to a mathematical problem selected from the problem type determination module (332), and the generated similar problem can be saved as a digital file in a format that allows text editing, such as PDF, HWP, or web text.

[0089] That is, the above-mentioned similar problem generation module (336) can generate a similar problem as shown in FIG. 3 (b) by applying numerical values ​​modified by the numerical transformation module (334) to the mathematical problem shown in FIG. 3 (a).

[0090] The above feedback providing unit (330) can create math problems selected by the problem type determination module (332) and similar problems generated by the similar problem generation module (336) in the form of a problem sheet and provide them to the learner through the learner terminal (400), etc., and the learner's problem-solving process can be captured by the first and second cameras (210, 220) provided in the data input unit and transmitted to the main server (300).

[0091] Next, the main server (300) may further include a communication unit (340), which is configured to perform wired and wireless communication processes with the database (100), the learner terminal (400), and the teacher terminal (500), and various known wired and wireless communication networks such as Ethernet, the Internet, WIFI, and LTE may be utilized.

[0092] Next, the learner terminal (400) may be a desktop computer, laptop computer, notebook, smartphone, tablet PC, mobile phone, etc. that is capable of communication and is carried by the learner.

[0093] The above learner terminal (400) is equipped with a communication function and can be connected to the main server (300) via a network, and can receive the learner's problem-solving analysis results and feedback provided therefrom provided by the main server (300), or receive the problem sheet and problem-solving provided by the main server (300) and simultaneously upload the problem sheet with completed solutions to the main server (300) in the form of an image file or a vector file.

[0094] Additionally, the mathematics learning system (10) according to the present invention may further include a teacher terminal (500), the teacher terminal (500) being a terminal possessed by teachers capable of presenting a model solution process for a mathematics problem, and may be a desktop computer, laptop computer, notebook, smartphone, tablet PC, mobile phone, etc. that is capable of communication.

[0095] That is, the math learning system (10) according to the present invention can analyze the strengths and weaknesses of learners in the problem-solving process, particularly weaknesses, by comparing the learners' problem-solving videos with the problem-solving videos of exemplary cases and analyzing the time taken for each step of the solution process, and can provide feedback that can compensate for the analyzed weaknesses along with training content, namely math problems, so that training is possible, and at the same time, learners can continuously improve their math problem-solving methods through the repetition of the above process.

[0097] Meanwhile, the mathematics learning method through deep analysis of mathematics problem-solving data according to the present invention relates to a mathematics learning method that can provide feedback to compensate for the shortcomings in the process of solving mathematics problems by learners using the aforementioned mathematics learning system (10), and its configuration may include, as shown in FIG. 4, a data collection step (S10), a data analysis step (S20), a comparison and evaluation step (S30), and a feedback provision step (S40).

[0098] First, the above data collection step (S10) is a process for collecting information including phenomena occurring during the process of learners solving math problems, namely problem-solving habits, solution strategies, solution processes, and time taken, and may be a process for acquiring video data of the learner solving problems captured by a first camera (210) installed in front of the learner and a second camera (220) installed above the learner.

[0099] Next, the data analysis step (S20) is a process for analyzing the problem-solving process of a learner captured by the first and second cameras (210, 220), which is the video data collected through the data collection step (S10). The data analysis unit (310) of the main server (300) can analyze the learner's problem-solving process or problem-solving habits by detecting the learner's eye movement and handwriting position during the problem-solving process through the collected video data.

[0100] To explain in more detail, the data analysis step (S20) may include a gaze analysis step (S22), a solution process analysis step (S24), and a first visualization step (S26). First, the gaze analysis step (S22) is a process for analyzing the movement of the learner's eyes during the problem-solving process using video data captured by the first camera (210). After identifying the learner's pupils in the video data captured by the first camera (210) through the gaze analysis module (312) provided in the data analysis unit (310) of the main server (300), the movement of the pupils is analyzed to estimate the direction in which the learner's gaze is directed, i.e., the gaze vector, thereby allowing the analysis of the learner's eye movement pattern, i.e., the change in gaze, during the problem-solving process.

[0101] Next, the above solution process analysis step (S24) may be a process for analyzing the problem-solving process by detecting the handwriting position in the problem-solving process using image data captured by the second camera (220).

[0102] That is, in the above-mentioned solution process analysis step (S24), the learner's handwriting position for problem solving detected by the solution process analysis module (314) of the above-mentioned data analysis unit (310) and the learner's pupil movement and gaze direction analyzed in the gaze analysis step (S22) are combined to divide the learner's problem solving process into four stages: problem reading stage, concept organization stage, calculation stage, and organization review stage, and the time required for each stage can be calculated through frame analysis of the above-mentioned video data.

[0103] Next, the first visualization step (S26) is a process for visualizing and displaying the content analyzed in the gaze analysis step (S22) and the solution process analysis step (S24) in the form of a chart or graph using the first visualization module (316) of the data analysis unit (310). The content visualized in the first visualization step (S26) can be displayed on a monitoring means (not shown) or transmitted to a learner terminal (400) via a network.

[0104] Next, the comparison and evaluation step (S30) is a process for deriving the strengths and weaknesses of the learner in the problem-solving process by comparing the analysis results from the data analysis step (S20), that is, the analysis results of the video data of the learner's problem-solving process, with exemplary cases in which exemplary problem-solving processes are filmed. The exemplary cases may be video data filmed using the first and second cameras (210, 220) of the data collection unit (200) of the problem-solving processes of high-achieving students or teachers.

[0105] The above comparison and evaluation step (S30) may include a solution strategy comparison step (S32), a calculation process comparison step (S34), a solution time comparison step (S36), and a second visualization step (S38). Since the process performed in the above steps and the technical configuration for performing the process may be the same as described in the comparison and evaluation unit (320) of the learning system (10) described above, a more detailed explanation thereof will be omitted.

[0106] Next, the feedback provision step (S40) is a process for providing feedback to the learner to compensate for weaknesses by considering the learner's strengths and weaknesses derived from the comparison and evaluation step (S30). In the feedback provision step (S40), the learner's inefficient problem-solving patterns are learned, and feedback that can improve the learner's problem-solving method through machine learning such as reinforcement learning can be continuously presented.

[0107] Additionally, in the feedback provision step (S40), individual customized training content may be provided by methods such as generating additional learning problems to compensate for the learner's weaknesses and providing them to the learner, and the process may include a problem type determination step (S42), a numerical transformation step (S44), and a similar problem generation step (S46).

[0108] That is, in the feedback providing unit (330) of the main server (300) as described above, math problems to compensate for the learner's weaknesses can be selected from the problem DB (110) of the database (100) through the problem type determination step (S42), and numeric information included in the problem conditions and question content of the selected math problems can be randomly transformed through the numeric transformation step (S44), and similar problems can be generated through the similar problem generation step (S46) by applying the numeric values ​​transformed by the numeric transformation step (S44) to the math problems selected in the problem type determination step (S42).

[0109] As described above, the selected math problems and generated similar problems are written in the form of a problem sheet and provided to the learner, and the learner's problem-solving process is captured by the first and second cameras (210, 220) and transmitted to the main server (300). Through the repetition of this process, the learner's weaknesses can be continuously improved.

[0110] Accordingly, the mathematics learning system and method through deep analysis of mathematics problem-solving data according to the present invention as described above has various advantages, such as analyzing a learner's mathematics problem-solving data to compare and evaluate the logical development method, concept application method, calculation accuracy, and solution time with a model solution to derive the learner's strengths and weaknesses, and automatically providing customized learning content for each learner based on the derived evaluation results, allowing learners to identify weaknesses in the problem-solving process and intuitively check the situation of improvement through feedback, thereby encouraging learners' motivation and stimulating their interest, while simultaneously improving learning efficiency and learning productivity.

[0112] The systems and methods described above may be implemented as hardware components, software components, and / or combinations of hardware and software components. For example, the components described in the embodiments may be implemented using one or more general-purpose or special-purpose computers, such as a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA), a programmable logic unit (PLU), a microprocessor, or any other device capable of executing and responding to instructions. The processing unit may execute an operating system (OS) and one or more software applications executed on said operating system. Additionally, the processing unit may access, store, manipulate, process, and generate data in response to the execution of the software. The processing unit may include multiple processors or one processor and one controller. Additionally, other processing configurations, such as parallel processors, are also possible.

[0113] Software may include computer programs, code, instructions, or a combination of one or more of these, and may configure a processing unit to operate as desired or command the processing unit independently or collectively. Software and / or data may be permanently or temporarily embodied in any type of machine, component, physical device, virtual equipment, computer storage medium or device, or transmitted signal wave in order to be interpreted by the processing unit or to provide instructions or data to the processing unit. Software may be distributed over networked computer systems and may be stored or executed in a distributed manner. Software and data may be stored on one or more computer-readable recording media.

[0114] The method according to the embodiment may be implemented in the form of program instructions that can be executed through various computer means and may be recorded on a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, etc., either alone or in combination. The program instructions recorded on the medium may be those specifically designed and configured for the embodiment, or they may be those known and available to a person with ordinary knowledge in the field of computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes; optical recording media such as CD-ROMs and DVDs; magneto-optical media such as floptical disks; and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, and flash memory.

[0115] Although the embodiments have been described above with reference to limited examples and drawings, those skilled in the art can make various modifications and variations from the description above. For example, suitable results can be achieved even if the described techniques are performed in a different order than described, and / or the components of the described system, structure, device, circuit, etc. are combined or assembled in a form different from described, or replaced or substituted by other components or equivalents. Explanation of the symbols

[0117] 10: Math Learning System 100 : Database 110 : Problem DB 120 : Best Practices DB 130 : Learner DB 200 : Data Collection Department 210 : 1st camera 220 : Second camera 300 : Main Server 310 : Data Analysis Department 312 : Eye-tracking module 314 : Solution Process Analysis Module 316 : 1st Visualization Module 320 : Comparison and Evaluation Section 322 : Solution Strategy Comparison Module 324 : Operation process comparison module 326 : Solution Time Comparison Module 328 : 2nd Visualization Module 330 : Feedback Provision Department 332 : Problem Type Determination Module 334 : Numerical transformation module 336 : Similar Problem Generation Module 340 : Communications Department 400 : Learner terminal 500 : Teacher Terminal S10: Data collection stage S20: Data Analysis Phase S22 : Eye gaze analysis stage S24 : Solution Process Analysis Stage S26: 1st Visualization Stage S30: Comparison and evaluation stage S32 : Solution Strategy Comparison Step S34: Comparison step of the operation process S36 : Solution time comparison stage S38: Second visualization stage S40: Feedback provision stage S42: Problem Type Determination Stage S44: Numerical transformation stage S46: Similar problem generation stage

Claims

Claim 1 A mathematics learning system through in-depth analysis of mathematics problem-solving data, comprising: a database that stores mathematics problems to be provided to learners, classified by unit, type, and difficulty level; a data collection unit that collects video data captured by the first and second cameras, including a first camera that captures the learner's problem-solving process from the front and a second camera that captures the learner's problem-solving process from the top; and a main server that is connected to the database and the data collection unit via a network, analyzes the video data collected through the data collection unit, and provides customized feedback for each learner. Claim 2 In claim 1, the main server comprises: a data analysis unit that analyzes image data collected from the data collection unit; a comparison and evaluation unit that derives the strengths and weaknesses of the learner in the problem-solving process by comparing the analysis results from the data analysis unit with best practices; and a feedback provision unit that provides feedback to compensate for weaknesses by considering the strengths and weaknesses of the learner derived from the comparison and evaluation unit; a mathematics learning system through in-depth analysis of mathematics problem-solving data. Claim 3 In claim 2, the data analysis unit comprises: an eye gaze analysis module that analyzes the learner's eye movement during the problem-solving process using video data captured by the first camera; a solution process analysis module that analyzes the problem-solving process by detecting the handwriting position during the problem-solving process using video data captured by the second camera; and a first visualization module that visualizes the analysis results by the eye gaze analysis module and the solution process analysis module; a mathematics learning system through in-depth analysis of mathematics problem-solving data. Claim 4 In claim 2, the comparison and evaluation unit comprises: a solution strategy comparison module that compares a learner's problem-solving strategy with a model case using the analysis results from the data analysis unit; a calculation process comparison module that compares a learner's calculation process with a model case using the analysis results from the data analysis unit; a solution time comparison module that compares the time taken for a learner to solve a problem with a model case using the analysis results from the data analysis unit; and a second visualization module that visualizes and displays the comparison results from the solution strategy comparison module, the calculation process comparison module, and the solution time comparison module; a mathematics learning system through in-depth analysis of mathematics problem-solving data. Claim 5 In claim 2, the feedback providing unit comprises: a problem type determination module that selects and provides a type of mathematical problem from a database that can compensate for the learner's weaknesses derived from the comparison and evaluation unit; a numerical transformation module that transforms numerical information included in the mathematical problem provided by the problem type determination module using AI functions; and a similar problem generation module that generates a similar problem by applying the numerical value transformed by the numerical transformation module to the mathematical problem selected by the problem type determination module. Claim 6 A method for learning mathematics through deep analysis of mathematics problem-solving data, comprising: a data collection step of acquiring video data capturing the problem-solving process of a learner using a first camera provided in front of the learner and a second camera provided above the learner; a data analysis step of analyzing the learner's problem-solving process through a data analysis unit of a main server using the video data collected in the data collection step; a comparison and evaluation step of deriving the learner's strengths and weaknesses in the problem-solving process by comparing the analysis results from the data analysis step with a model case in which a model problem-solving process is captured; and a feedback provision step of providing feedback to supplement weaknesses by considering the learner's strengths and weaknesses derived in the comparison and evaluation step.

Images