Online precision teaching and tutoring method and system based on artificial intelligence and big data

Abstract

The present invention relates to the field of teaching and tutoring technology, and specifically, to an online precision teaching and tutoring method based on artificial intelligence and big data, and a system and storage medium thereof. It includes a client, a wrong question collection unit, a question type integration unit, an online comment unit, and an interactive discussion unit. The present invention is designed to formulate teaching question types using a precise question type synthesis algorithm for the summarized wrong questions, publish teaching tasks to individual student clients, and make targeted teaching plans based on individual student situations, which is conducive to precise teaching and tutoring for individual students' weak question types, and is more practical. Students can selectively view video analysis of their own learning errors, and can also view video analysis of question types that they do not understand well but have answered correctly, which is conducive to students' more thorough understanding of question type problem-solving methods, and targeted teaching planning for students' learning, which is conducive to learning progress.

Description

Technical Field

[0001] The present invention relates to the field of teaching and tutoring technology, and in particular to an online precise teaching and tutoring method based on artificial intelligence and big data, and a system and storage medium thereof. Background Art

[0002] With the application of information technology, mobile communication technology and big data technology in real life, we are facing changes and challenges in teaching forms and methods. In line with the trend of the "Internet Education" era and with the implementation of the "Internet +" online classroom project, the basic goal of sharing high-quality resources in the region has been achieved. Rural weak schools obtain high-quality educational resources through remote online interaction. However, for the teachers in the classroom, it is still limited to understand the learning data of remote students only through the big screen, and communication and exchange are still limited. Targeted teaching and guidance cannot be provided according to the weak questions of different students, resulting in the teacher's guidance not being targeted, which not only wastes students' time, but also leads to the lack of guidance on questions they do not understand, which is not conducive to rapid progress. Summary of the invention

[0003] The purpose of the present invention is to provide an online precision teaching and tutoring method based on artificial intelligence and big data, and its system and storage medium to solve the problems raised in the above-mentioned background technology.

[0004] In order to solve the above technical problems, one of the purposes of the present invention is to provide an online precision teaching and tutoring method based on artificial intelligence and big data, comprising the following steps:

[0005] S1. Students take photos on the student client to identify wrong questions and store the wrong questions.

[0006] S2, extract the easy-to-make mistakes questions from big data through web crawlers, and summarize the easy-to-make mistakes questions from big data and the students’ personal mistakes;

[0007] S3. Use the accurate question type synthesis algorithm to formulate teaching question types based on the summarized wrong questions, publish the teaching tasks to the student's personal client, and make targeted teaching plans based on the student's personal situation, which is conducive to accurate teaching and guidance for the student's weak question types, and is more practical;

[0008] S4. After the student completes the teaching plan, the teacher reviews and comments on the student's learning situation on the teacher client, and conducts video analysis of the teaching question type. The student can choose to play the video teaching of any teaching question type on the student client, which is convenient for the student to view the video analysis of the mistakes in their learning. The student can also view the video analysis of the question type that they do not understand well but answered correctly, which is conducive to the student's more thorough understanding of the question type and problem-solving method;

[0009] S5. The student client uses a queuing interaction algorithm to transmit the question types that the student client does not understand to the teacher client, so that the teacher client and the student client can have a one-to-one interactive discussion;

[0010] S6. Generate personalized reports on student learning on the teacher client and update the learning progress to the student client, so as to facilitate targeted update planning for student learning and promote learning progress.

[0011] As a further improvement of the technical solution, the photographing and identifying wrong questions adopts an image recognition algorithm, including the following steps:

[0012] Image preprocessing, enhancing the image through denoising and image regularization;

[0013] Image invariant moment feature extraction, extracting 7 invariant moments of the image;

[0014] Image recognition uses a neural network algorithm to attribute the errors in the network output to the "faults" of each connection weight. The errors in the output layer are propagated back to the input layer layer by layer to "share" them among the units in each layer, thereby obtaining the reference errors of the units in each layer in order to adjust the corresponding connection weights, and adaptively adjust the weights to achieve satisfactory output content.

[0015] By using the invariant moment features, 7 invariant moments of the image are extracted and combined with the neural network algorithm for image target recognition. This algorithm has less computational complexity and is easy to implement. It also avoids the influence of the change of the scale factor on the invariant moment. It is a very effective method for the recognition of small sun targets in images.

[0016] As a further improvement of the technical solution, the web crawler comprises the following steps:

[0017] First, select a portion of carefully selected seed URLs;

[0018] Put these URLs into the queue of URLs to be crawled;

[0019] Take the URL to be crawled from the queue of URLs to be crawled, resolve the DNS, and get the host's IP, and download the web page corresponding to the URL and store it in the downloaded web page library. In addition, put these URLs into the queue of crawled URLs;

[0020] Analyze the URLs in the crawled URL queue, analyze other URLs in it, and put the URLs into the queue of URLs to be crawled, so as to enter the next cycle. It is possible to extract more comprehensive error-prone questions based on big data.

[0021] As a further improvement of the present technical solution, the precise question type synthesis algorithm includes the following steps:

[0022] Extract keywords from students' individual wrong questions and big data's easy-to-make-mistake questions respectively;

[0023] Define wrong questions with the same keyword as one question type;

[0024] Integrate various question types to form a wrong question type library;

[0025] Therefore, if the keywords extracted from the student's personal wrong question content match the keywords of the easy-to-make mistakes content in big data, the two questions will be defined as the same question type. If the student's personal wrong question content does not match the keywords of the easy-to-make mistakes content in big data, the student's personal wrong question will be defined as a separate question type. On the contrary, if the keywords of the easy-to-make mistakes content in big data do not match the student's personal wrong question content keywords, the easy-to-make mistakes content in big data will be defined as a separate question type. This will continue until all question types are classified and stored in the wrong question type library, so that targeted tutoring can be provided during subsequent learning.

[0026] As a further improvement of the technical solution, the precise question type synthesis algorithm includes the following calculation formula:

[0027] Student's personal collection of wrong questions:

[0028] A={a1, a2, ..., a k}

[0029] A collection of easy-to-error questions about big data:

[0030] B={b1, b2, ..., b k}

[0031] Wrong question collection:

[0032] E=A+B={(a+b)1, (a+b)2,..., (a+b) k}

[0033] Among them, A is the set of wrong questions of individual students, B is the set of easy-to-make mistakes in big data, and E is the total set of wrong questions. a1, a2, ..., a k For a single student's personal wrong question type, b1, b2, ..., b k For single big data, easy-to-error questions, (a+b)1, (a+b)2, ..., (a+b) k In order to integrate students' individual wrong question types and big data's easy-to-make mistakes question types into a single general question type, all question types can be summarized and summarized to obtain a more complete wrong question type library, which is conducive to accurate tutoring and teaching.

[0034] As a further improvement of the technical solution, the video analysis adopts a video denoising algorithm to improve the clarity of remote video recording. Noise filtering is likely to filter out normal image information. Therefore, while performing noise filtering, the normal image information is accurately protected, including the following steps:

[0035] Classify the noisy images into noisy images and non-noisy images;

[0036] Determine the reasonable distribution of information in the noisy image, perform effective filtering output, and retain the maximum image information. If the noise contained in the image does not form effective information, then filter it directly and output it directly without any processing;

[0037] As a further improvement of the technical solution, the video denoising algorithm includes the following formula:

[0038] Additive noise:

[0039] g=f+n

[0040] Additive noise is unrelated to the image signal strength. For example, the "channel noise" introduced during the image transmission process is the noise of the camera scanning the image.

[0041] Multiplicative noise:

[0042] g=f+fn

[0043] Among them, g is the image with noise, f is the ideal noise-free image, and n is the sum of noise. Multiplicative noise is related to image signal and often changes with the change of image signal, such as the noise in flying spot scanning image, TV scanning raster, film grain noise, etc. The relationship between this kind of noise and image can be described by the above formula;

[0044] Quantization noise:

[0045]

[0046] Among them, G is the quantization noise parameter, n1 is the number of pixels of the digital image, m is the number of pixels of the original image, L is the transmission distance of the educational video image, and v is the data transmission speed. Quantization noise is the main noise source of digital images, and its size shows the difference between digital images and original images. The best way to reduce this noise is to use the gray level probability density formula to optimize data processing;

[0047] Salt and pepper noise:

[0048]

[0049] Among them, R is the salt and pepper noise parameter, x is the number of pixels of the initial image, y is the number of regions after the initial image is segmented, p is the number of white dots in the initial image, and q is the number of black dots in the initial image. This type of noise is caused by image cutting, that is, white dots on black images, black dots on white images, errors introduced in the transform domain, and transformation noise caused by inverse transformation of the image.

[0050] As a further improvement of the technical solution, the queuing interaction algorithm includes the following steps:

[0051] The student client extracts the question types that it does not understand and transmits them to the teacher client;

[0052] Question types are queued in the waiting queue according to priority;

[0053] When the teacher's client's service desk is idle, select the question type with the highest priority from the waiting queue;

[0054] The teacher checks the question types on the teacher client and communicates with the student client one-on-one;

[0055] The teacher can receive multiple questions sent by students from the student client in turn on the teacher client, which is convenient for one-on-one teaching and enables students to understand more thoroughly. Through the queuing interaction algorithm, it can avoid confusion caused by multiple students communicating at the same time, making the teaching and tutoring methods more orderly.

[0056] The second object of the present invention is to provide an online precision teaching and tutoring system based on artificial intelligence and big data, including any one of the above-mentioned online precision teaching and tutoring methods based on artificial intelligence and big data, including a client, a wrong question collection unit, a question type integration unit, an online comment unit and an interactive discussion unit, the client including a student client and a teacher client, the client is designed based on the underlying architecture of Linux and Window systems, the wrong question collection unit is used to store and collect students' personal wrong questions and easy-to-make mistakes in big data, the question type integration unit is used to integrate the wrong questions collected by the wrong question collection unit into the same question type to generate teaching question types, the online comment unit is used for teachers to provide video tutoring for teaching question types completed by students, and the interactive discussion unit is used to achieve one-to-one communication between teachers and students.

[0057] The third object of the present invention is to provide a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements any of the above-mentioned online precision teaching and tutoring methods based on artificial intelligence and big data.

[0058] Compared with the prior art, the present invention has the following beneficial effects:

[0059] 1. In the online precision teaching and tutoring method based on artificial intelligence and big data and its system and storage medium, the summarized wrong questions are formulated into teaching question types using a precise question type synthesis algorithm, and the teaching tasks are released to the student's personal client. Targeted teaching plans are made based on the student's personal situation, which is conducive to precise teaching and tutoring for students' individual weak question types. It is more practical. Students can selectively view the video analysis of their own learning errors, or view the video analysis of question types that they do not understand well but have answered correctly, which is conducive to students' more thorough understanding of question type solving methods, targeted teaching plans for students' learning, and conducive to learning progress.

[0060] 2. In the online precision teaching and tutoring method based on artificial intelligence and big data and its system and storage medium, the video denoising algorithm is adopted through video analysis to improve the clarity of remote video recording. Noise filtering is likely to filter out normal image information. Therefore, while performing noise filtering, normal image information is accurately protected.

[0061] 3. In the online precision teaching and tutoring method based on artificial intelligence and big data and its system and storage medium, a queuing interaction algorithm is used to enable the teacher to receive question types sent by multiple students from the student client in turn on the teacher client, which is convenient for one-to-one teaching and enables students to understand more thoroughly. The queuing interaction algorithm can avoid confusion caused by multiple students communicating at the same time, making the teaching and tutoring method more orderly. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 This is a schematic diagram of the overall architecture of Example 1;

[0063] Figure 2 This is a block diagram of the queue interaction algorithm of Example 1. DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] Example 1

[0066] like Figure 1-2 As shown, one of the purposes of this embodiment is to provide an online precision teaching and tutoring method based on artificial intelligence and big data, comprising the following steps:

[0067] S1. Students take photos on the student client to identify wrong questions and store the wrong questions.

[0068] In this embodiment, the photographing and identifying wrong questions adopts an image recognition algorithm, including the following steps:

[0069] Image preprocessing, enhancing the image through denoising and image regularization;

[0070] Image invariant moment feature extraction, extracting 7 invariant moments of the image;

[0071] Image recognition uses a neural network algorithm to attribute the errors in the network output to the "faults" of each connection weight. The errors in the output layer are propagated back to the input layer layer by layer to "share" them among the units in each layer, thereby obtaining the reference errors of the units in each layer in order to adjust the corresponding connection weights, and adaptively adjust the weights to achieve satisfactory output content.

[0072] By using the invariant moment features, 7 invariant moments of the image are extracted and combined with the neural network algorithm for image target recognition. This algorithm has less computational complexity and is easy to implement. It also avoids the influence of the change of the scale factor on the invariant moment. It is a very effective method for the recognition of small sun targets in images.

[0073] S2, extract the easy-to-make mistakes questions from big data through web crawlers, and summarize the easy-to-make mistakes questions from big data and the students’ personal mistakes;

[0074] Specifically, the web crawler includes the following steps:

[0075] First, select a portion of carefully selected seed URLs;

[0076] Put these URLs into the queue of URLs to be crawled;

[0077] Take the URL to be crawled from the queue of URLs to be crawled, resolve the DNS, and get the host's IP, and download the web page corresponding to the URL and store it in the downloaded web page library. In addition, put these URLs into the queue of crawled URLs;

[0078] Analyze the URLs in the crawled URL queue, analyze other URLs in it, and put the URLs into the queue of URLs to be crawled, so as to enter the next cycle. It is possible to extract more comprehensive error-prone questions based on big data.

[0079] S3. Use the accurate question type synthesis algorithm to formulate teaching question types based on the summarized wrong questions, publish the teaching tasks to the student's personal client, and make targeted teaching plans based on the student's personal situation, which is conducive to accurate teaching and guidance for the student's weak question types, and is more practical;

[0080] In order to facilitate subsequent learning and provide targeted tutoring, the precise question type synthesis algorithm includes the following steps:

[0081] Extract keywords from students' individual wrong questions and big data's easy-to-make-mistake questions respectively;

[0082] Define wrong questions with the same keyword as one question type;

[0083] Integrate various question types to form a wrong question type library;

[0084] Therefore, if the keywords extracted from the student's personal wrong question content match the keywords of the easy-to-make mistakes content in big data, the two questions will be defined as the same question type. If the student's personal wrong question content does not match the keywords of the easy-to-make mistakes content in big data, the student's personal wrong question will be defined as a separate question type. On the contrary, if the keywords of the easy-to-make mistakes content in big data do not match the student's personal wrong question content keywords, the easy-to-make mistakes content in big data will be defined as a separate question type. This will continue until all question types are classified and stored in the wrong question type library, so that targeted tutoring can be provided during subsequent learning.

[0085] Specifically, the precise question type synthesis algorithm includes the following calculation formula:

[0086] Student's personal collection of wrong questions:

[0087] A={a1, a2, ..., a k}

[0088] A collection of easy-to-error questions about big data:

[0089] B={b1, b2, ..., b k}

[0090] Wrong question collection:

[0091] E=A+B={(a+b)1, (a+b)2,..., (a+b) k}

[0092] Among them, A is the set of wrong questions of individual students, B is the set of easy-to-make mistakes in big data, and E is the total set of wrong questions. a1, a2, ..., a k For a single student's personal wrong question type, b1, b2, ..., b k For single big data, easy-to-error questions, (a+b)1, (a+b)2, ..., (a+b) k In order to integrate students' individual wrong question types and big data's easy-to-make mistakes question types into a single general question type, all question types can be summarized and summarized to obtain a more complete wrong question type library, which is conducive to accurate tutoring and teaching.

[0093] S4. After the student completes the teaching plan, the teacher reviews and comments on the student's learning situation on the teacher client, and conducts video analysis of the teaching question type. The student can choose to play the video teaching of any teaching question type on the student client, which is convenient for the student to view the video analysis of the mistakes in their learning. The student can also view the video analysis of the question type that they do not understand well but answered correctly, which is conducive to the student's more thorough understanding of the question type and problem-solving method;

[0094] In order to improve the clarity of video teaching, the video analysis adopts a video denoising algorithm, which can improve the clarity of remote video recording. Noise filtering is likely to filter out normal image information. Therefore, while performing noise filtering, the normal image information is accurately protected, including the following steps:

[0095] Classify the noisy images into noisy images and non-noisy images;

[0096] Determine the reasonable distribution of information in the noisy image, perform effective filtering output, and retain the maximum image information. If the noise contained in the image does not form effective information, then filter it directly and output it directly without any processing;

[0097] Specifically, the video denoising algorithm includes the following formula:

[0098] Additive noise:

[0099] g=f+n

[0100] Additive noise is unrelated to the image signal strength. For example, the "channel noise" introduced during the image transmission process is the noise of the camera scanning the image.

[0101] Multiplicative noise:

[0102] g=f+fn

[0103] Among them, g is the image with noise, f is the ideal noise-free image, and n is the sum of noise. Multiplicative noise is related to image signal and often changes with the change of image signal, such as the noise in flying spot scanning image, TV scanning raster, film grain noise, etc. The relationship between this kind of noise and image can be described by the above formula;

[0104] Quantization noise:

[0105]

[0106] Among them, G is the quantization noise parameter, n1 is the number of pixels of the digital image, m is the number of pixels of the original image, L is the transmission distance of the educational video image, and v is the data transmission speed. Quantization noise is the main noise source of digital images, and its size shows the difference between digital images and original images. The best way to reduce this noise is to use the gray level probability density formula to optimize data processing;

[0107] Salt and pepper noise:

[0108]

[0109] Among them, R is the salt and pepper noise parameter, x is the number of pixels of the initial image, y is the number of regions after the initial image is segmented, p is the number of white dots in the initial image, and q is the number of black dots in the initial image. This type of noise is caused by image cutting, that is, white dots on black images, black dots on white images, errors introduced in the transform domain, and transformation noise caused by inverse transformation of the image.

[0110] S5. The student client uses a queuing interaction algorithm to transmit the question types that the student client does not understand to the teacher client, so that the teacher client and the student client can have a one-to-one interactive discussion;

[0111] In order to improve the order of communication, Figure 2 As shown, the queue interaction algorithm includes the following steps:

[0112] The student client extracts the question types that it does not understand and transmits them to the teacher client;

[0113] Question types are queued in the waiting queue according to priority;

[0114] When the teacher's client's service desk is idle, select the question type with the highest priority from the waiting queue;

[0115] The teacher checks the question types on the teacher client and communicates with the student client one-on-one;

[0116] The teacher can receive multiple questions sent by students from the student client in turn on the teacher client, which is convenient for one-on-one teaching and enables students to understand more thoroughly. Through the queuing interaction algorithm, it can avoid confusion caused by multiple students communicating at the same time, making the teaching and tutoring methods more orderly.

[0117] S6. Generate personalized reports on student learning on the teacher client and update the learning progress to the student client, so as to facilitate targeted update planning for student learning and promote learning progress.

[0118] The second object of the present invention is to provide an online precision teaching and tutoring system based on artificial intelligence and big data, including any one of the above-mentioned online precision teaching and tutoring methods based on artificial intelligence and big data, including a client, a wrong question collection unit, a question type integration unit, an online comment unit and an interactive discussion unit, the client including a student client and a teacher client, the client is designed based on the underlying architecture of Linux and Window systems, the wrong question collection unit is used to store and collect students' personal wrong questions and easy-to-make mistakes in big data, the question type integration unit is used to integrate the wrong questions collected by the wrong question collection unit into the same question type to generate teaching question types, the online comment unit is used for teachers to provide video tutoring on teaching question types completed by students, and the interactive discussion unit is used to achieve one-to-one communication between teachers and students.

[0119] The third object of the present invention is to provide a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements any of the above-mentioned online precision teaching and tutoring methods based on artificial intelligence and big data.

[0120] Optionally, the storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0121] Optionally, the present invention also provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the above-mentioned online precision teaching and tutoring methods based on artificial intelligence and big data.

[0122] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.

[0123] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and descriptions are only preferred examples of the present invention and are not intended to limit the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. An online precision teaching and tutoring method based on artificial intelligence and big data, characterized by: The following steps are involved: S1. Students take photos on the student client to identify wrong questions and store the wrong questions. The method of taking photos to identify wrong questions adopts an image recognition algorithm, including the following steps: Image preprocessing, enhancing the image through denoising and image regularization; Image invariant moment feature extraction, extracting 7 invariant moments of the image; Image recognition uses a neural network algorithm to attribute the errors in the network output to the "faults" of each connection weight. The errors in the output layer are propagated back to the input layer layer by layer to "share" them among the units in each layer, thereby obtaining the reference errors of the units in each layer in order to adjust the corresponding connection weights, and adaptively adjust the weights to achieve satisfactory output content. S2, extract the easy-to-make mistakes questions from big data through web crawlers, and summarize the easy-to-make mistakes questions from big data and the students’ personal mistakes; Specifically, the web crawler includes the following steps: First, select a portion of carefully selected seed URLs; Put these URLs into the queue of URLs to be crawled; Take the URL to be crawled from the queue of URLs to be crawled, resolve the DNS, and get the host's IP, and download the web page corresponding to the URL and store it in the downloaded web page library. In addition, put these URLs into the queue of crawled URLs; Analyze the URLs in the crawled URL queue, analyze other URLs in it, and put the URLs into the queue of URLs to be crawled, so as to enter the next cycle and extract more comprehensive error-prone questions based on big data; S3. Use the accurate question type synthesis algorithm to formulate teaching question types based on the summarized wrong questions, publish the teaching tasks to the student's personal client, and make targeted teaching plans based on the student's personal situation, which is conducive to accurate teaching and guidance for the student's weak question types, and is more practical; In order to facilitate targeted tutoring during subsequent learning, the precise question type synthesis algorithm includes the following steps: Extract keywords from students' individual wrong questions and big data's easy-to-make-mistake questions respectively; Define wrong questions with the same keyword as one question type; Integrate various question types to form a wrong question type library; Therefore, if the keywords of the students' personal wrong questions are matched with the keywords of the easy-to-make mistakes in the big data, the two questions are defined as the same question type. If the students' personal wrong questions do not match the keywords of the easy-to-make mistakes in the big data, the students' personal wrong questions are defined as a separate question type. On the contrary, if the keywords of the easy-to-make mistakes in the big data do not match the keywords of the students' personal wrong questions, the easy-to-make mistakes in the big data are defined as a separate question type. Until all question types are classified and stored in the wrong question type library, it is convenient for targeted tutoring during subsequent learning. Specifically, the precise question type synthesis algorithm includes the following calculation formula: Student's personal collection of wrong questions: <h2 style=";text-align:left;direction:ltr">A={a1, a2,..., a<h2 style=";text-align:left;direction:ltr"> k <h2 style=";text-align:left;direction:ltr">} A collection of easy-to-error questions about big data: B={b1、b2、...、b k } Wrong question collection: E=A+B={(a+b)1、(a+b)2、...、(a+b) k } Among them, A is the set of wrong questions of individual students, B is the set of easy-to-make mistakes in big data, and E is the total set of wrong questions. a1, a2, ..., a k For a single student's personal wrong question type, b1, b2, ..., b k For single big data, easy-to-error questions: (a+b)1, (a+b)2, ..., (a+b) k In order to integrate students' individual wrong question types and big data's easy-to-make-mistake question types into a single general question type, all question types are summarized and summarized to obtain a more complete wrong question type library, which is conducive to accurate tutoring and teaching; S4. After the student completes the teaching plan, the teacher reviews and comments on the student's learning situation on the teacher client, and provides video analysis of the teaching question type. The student can choose to play the video teaching of any teaching question type on the student client; this also includes viewing the video analysis of the question type that is not well understood but answered correctly; The video analysis adopts a video denoising algorithm, including the following steps: Classify the noisy images into noisy images and non-noisy images; Determine the reasonable distribution of information in noisy images, perform effective filtering output, and retain the maximum image information; The video denoising algorithm includes the following formula: Additive noise: g=f+n Multiplicative noise: g=f+fn Among them, g is the noisy image, f is the ideal noise-free image, and n is the sum of the noise; Quantization noise: Among them, G is the quantization noise parameter, n1 is the number of pixels of the digital image, m is the number of pixels of the original image, L is the transmission distance of the educational video image, and v is the data transmission speed; Salt and pepper noise: Where R is the salt and pepper noise parameter, x is the number of pixels in the initial image, y is the number of regions after segmenting the initial image, p is the number of white dots in the initial image, and q is the number of black dots in the initial image; S5. The student client uses a queuing interaction algorithm to transmit the question types that the student client does not understand to the teacher client, so that the teacher client and the student client can have a one-to-one interactive discussion; The queue interaction algorithm comprises the following steps: The student client extracts the question types that it does not understand and transmits them to the teacher client; Question types are queued in the waiting queue according to priority; When the teacher's client's service desk is idle, select the question type with the highest priority from the waiting queue; The teacher checks the question types on the teacher client and communicates with the student client one-on-one; S6. Generate personalized reports on student learning on the teacher client and update the learning progress on the student client.

2. An online precision teaching and tutoring system based on artificial intelligence and big data, comprising the online precision teaching and tutoring method based on artificial intelligence and big data as claimed in claim 1, characterized in that: It includes a client, a wrong question collection unit, a question type integration unit, an online comment unit and an interactive discussion unit. The client includes a student client and a teacher client. The wrong question collection unit is used to store and collect students' personal wrong questions and easy-to-make mistakes in big data. The question type integration unit is used to integrate the wrong questions collected by the wrong question collection unit into the same question type to generate teaching question types. The online comment unit is used for teachers to provide video tutoring for teaching questions completed by students. The interactive discussion unit is used to achieve one-on-one communication between teachers and students.

3. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by the processor, the online precision teaching and tutoring method based on artificial intelligence and big data as claimed in claim 1 is implemented.

Images