An AI-based college course selection recommendation system

Abstract

This invention discloses an artificial intelligence-based college course selection recommendation system, belonging to the field of course selection recommendation technology. The invention generates historical tag information by collecting student course selection information and matching it with matching feature information. It then analyzes and determines dynamic preference values ​​based on a large amount of historical tag information. Through the analysis of these dynamic preference values, it constructs a first, second, and third course selection recommendation set. The system extracts the top three courses from the first course selection recommendation set based on their dynamic preference values ​​and displays them prominently on the page. When a student registers and logs in, the system automatically matches them with students in the same major and recommends courses based on their major preferences, thereby enhancing the correlation between the preferences of students in the same major. Furthermore, it uses the matching system of the second and third course selection recommendation sets and the recommendation situation of the current year to correlate the non-major preferences of students in the same major.

Description

Technical Field

[0001] This invention relates to the field of college course selection recommendation technology, and in particular to a college course selection recommendation system based on artificial intelligence. Background Technology

[0002] With the construction of campus networks in Chinese universities, the development of internet-based application systems is booming and playing a significant role. For example, many universities in my country have online enrollment systems, various management information systems, course selection systems, and even online clinics and remote diagnostic systems developed by some medical colleges. These are all application systems based on campus networks. However, in the existing course selection systems in higher education institutions, courses are divided into general education courses, public basic courses, professional basic courses, and professional courses. Recommendations are usually made based on students' scores after exams. This makes it impossible to make targeted intelligent course matching recommendations for students in the same major and year based on students' self-selection information. On the other hand, it also makes synchronous recommendations for courses outside the student's major, resulting in a low level of intelligent recommendation.

[0003] To address the aforementioned technical shortcomings, a solution is proposed. Summary of the Invention

[0004] The purpose of this invention is to: on the one hand, intelligently match and recommend courses to students of the same major and year, and on the other hand, synchronously recommend courses to students of other majors.

[0005] To achieve the above objectives, the present invention adopts the following technical solution:

[0006] An artificial intelligence-based college course selection recommendation system includes a visualization window and a course selection recommendation platform. The visualization window and the course selection recommendation platform are connected by a signal via the Internet, and the visualization window includes a registration and login unit and a data collection and transmission unit.

[0007] The registration and login unit is used by students to register by submitting their information through a visual window. The unit marks the successful registration time and sends the registered student information to the course selection recommendation platform for storage. The student information includes the student's name, age, ID number, mobile phone number, grade, department, and major. The grade and major are marked as matching feature information. The collection and sending unit is used to collect students' course selection information and send it to the course selection recommendation platform for storage.

[0008] The course selection recommendation platform receives students' course selection information and matches it with student information to generate historical tag information. Historical tag information with the same matching feature information is stored in the same area and packaged to generate corresponding data packages to be analyzed. The data in the data packages to be analyzed is analyzed and judged to generate dynamic preference values. The course selection recommendation set is constructed through the analysis of dynamic preference values, and the course selection recommendation set is sent to the visualization window to refresh and display the corresponding courses in real time.

[0009] Furthermore, the specific steps for generating preference dynamic value analysis are as follows:

[0010] The system obtains the storage volume and / or storage time of the data packet to be analyzed. When either the storage volume or the storage time of the data packet to be analyzed is greater than or equal to the corresponding preset limit, the data packet to be analyzed is decompressed. The number of courses selected is then extracted, and courses with a selection volume less than the preset number are removed. The selection volume of the removed courses is marked as the remaining course selection volume Qi, where i ranges from 1 to n, where n is a positive integer, and Q1, Q2, ..., Qn correspond to the remaining course selection volumes. The average selection volume and standard deviation of the remaining courses are then calculated using formulas. The preference dispersion value is obtained based on the ratio of the standard deviation to the average selection volume. The preference dispersion value is then multiplied by the corresponding remaining course selection volume to obtain the median value for each course. Finally, several real-time median values ​​within the most recent preset time period are extracted and averaged to obtain the dynamic preference value.

[0011] Furthermore, the analysis and construction process of the course selection recommendation set includes:

[0012] Extract the dynamic preference values ​​of courses within the same major, then sort the dynamic preference values ​​from largest to smallest, and use the dynamic preference values ​​to sort the corresponding courses to generate the first set of recommended courses;

[0013] Obtain multiple priority course selection recommendation sets under the department, extract the weighting parameter one corresponding to the sorting, multiply the weighting parameter one by the preference dynamic value corresponding to the sorting, and then average the results of multiple multiplications to generate multiple first preference mean values. Sort the first preference mean values, and then form a second course selection recommendation set by combining the sorted first preference mean values ​​with the corresponding courses.

[0014] Obtain the complete set of priority course recommendations, extract the weighting parameter 2 corresponding to each sort, multiply the weighting parameter 2 by the corresponding dynamic preference value, average the results of multiple multiplications to generate multiple second preference mean values, sort the second preference mean values, and then form a third course recommendation set by combining the sorted second preference mean values ​​with the corresponding courses.

[0015] Furthermore, the course selection recommendation platform has a learning acquisition unit and an integration feedback unit connected by signals, and the feedback monitoring unit has a division and adjustment unit connected by signals.

[0016] During the process of students selecting and learning courses through a visual window, the learning acquisition unit is used to collect visual learning images and learning information of students during the course learning process within a preset time, and sends the visual learning images and learning information of students during the course learning process to the feedback monitoring unit.

[0017] The integrated feedback unit is used to receive visual learning images within a preset time, analyze and process them to generate visual feedback factors, determine the generation of learning reminder signals based on the visual feedback factors, edit the learning reminder text after the learning reminder signal is generated, and send the learning reminder text to the visualization window interface for top scrolling.

[0018] It also obtains the visual feedback factor during the entire course learning cycle and averages it to obtain the feedback mean. Then, it normalizes the data in the learning information and the feedback mean and combines them to generate the course's ability correction value. Then, it multiplies the course's ability correction value and its corresponding preference dynamic value with the weight parameter two, and adds the results of the multiplication to obtain a new preference dynamic value. The new preference dynamic value is then used to sort the courses from largest to smallest and refresh the order of the corresponding courses in the course selection recommendation set.

[0019] Furthermore, the specific generation process of the visual feedback factor is as follows:

[0020] Extract the student's head contour from the visual learning image. Mark the central axis of the visualization window as the focusing line and the midpoint of the focusing line as the focal point. Using the midpoint between the two eyes in the head contour as the dynamic point, construct a line perpendicular to the focusing line, with the perpendicular line placed at the focal point. Based on the perpendicular line, draw parallel lines to the dynamic point, extending from the focal point to intersect the parallel lines perpendicularly to form an intersection point. Connect the focal point, dynamic point, and intersection point end-to-end to form a virtual perspective triangle. Collect the angle between the line connecting the focal point and dynamic point in the virtual perspective triangle and the parallel line, and record the absolute value of the angle as the virtual perspective. Obtain the distance between the focal point and dynamic point and mark it as the virtual distance. Multiply the virtual perspective, virtual distance, and weight parameter one, and add the results to obtain the visual feedback factor. When the visual feedback factor is greater than the preset feedback value, a learning reminder signal is generated and stored. When the visual feedback factor is less than the preset feedback value, no learning reminder signal is generated.

[0021] Furthermore, the normalization process for the recommended correction values ​​is as follows:

[0022] The visual feedback factors throughout the entire course learning cycle are obtained and averaged to obtain the feedback mean; then, a course completion score matrix is ​​constructed using learning information; the matching degree between the students to be analyzed and the students in the matrix is ​​obtained, and the matching degree of different courses in this major is extracted and averaged to obtain the matching mean; the feedback mean and the matching mean are normalized to obtain the course ability correction value.

[0023] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0024] (1) This invention stores student information through student registration and marks the grade and major within the student information as matching feature information. It is also used to collect student course selection information and match it with the matching feature information to generate historical tag information. Through the analysis of a large amount of historical tag information, a dynamic preference value is generated. The first course selection recommendation set, the second course selection recommendation set, and the third course selection recommendation set are constructed through the analysis of the dynamic preference value. The top three courses in the first course selection recommendation set are extracted and scrolled on the top of the page. When a student registers and logs in, the system automatically matches students of the same major and recommends them to them based on their professional preferences, so as to enhance the correlation of the preferences of students of the same major. The system also uses the matching system of the second and third course selection recommendation sets and the recommendation situation of the current grade to associate the non-major preferences of students of the same major. The system combines professional and student preferences for recommendation, making it more professional and intelligent. On the one hand, it performs intelligent course matching recommendation for students of the same major and the same grade, and on the other hand, it performs synchronous recommendation for courses of non-major, thereby improving the intelligence of the recommendation.

[0025] (2) This invention also collects visual learning images and learning information of students during the course learning process within a preset time. It generates visual feedback factors by analyzing and processing the visual learning images, and generates learning reminder signals by judging the visual feedback factors. After the learning reminder signal is generated, the learning reminder text is edited and sent to the visualization window interface for scrolling at the top to remind students to pay attention. It also normalizes and combines the learning information and feedback average to generate the course ability correction value. Then, it combines the course ability correction value with the corresponding preference dynamic value to generate a new preference dynamic value. The new preference dynamic value is used to sort from large to small and refresh the order of corresponding courses in the course selection recommendation set to further improve the course recommendation. On the one hand, it realizes the reminder of students' attention during the learning process. On the other hand, it combines students' learning ability with students' preferred courses to quantify, improve the accuracy of course selection recommendation, avoid students blindly selecting courses, and thus further improve the intelligence of this system. Attached Figure Description

[0026] Figure 1 A first structural flowchart of the present invention is shown;

[0027] Figure 2 A flowchart of the second structure of the present invention is shown; Detailed Implementation

[0028] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] Example 1:

[0030] like Figure 1 As shown, a college course selection recommendation system based on artificial intelligence includes a visualization window and a course selection recommendation platform. The visualization window and the course selection recommendation platform are connected by a signal via the Internet, and the visualization window includes a registration and login unit and a data collection and transmission unit.

[0031] The registration and login unit allows students to register by submitting their information through a visual window. The unit marks the successful registration time and sends the registered student information to the course selection recommendation platform for storage. This student information includes name, age, ID number, mobile phone number, grade, department, and major. Grade and major are marked as matching features. The collection and sending unit collects student course selection information and sends it to the course selection recommendation platform for storage. Students in the same major and grade are categorized using the matching features to ensure a benchmark for subsequent course recommendations.

[0032] The course selection recommendation platform is used to receive students' course selection information, match this information with student information to generate historical tag information, and store historical tag information with the same matching feature information in the same area and package them to generate corresponding data packages to be analyzed.

[0033] Obtain the amount of stored data and / or the storage time of the data packet to be analyzed. The unit of the amount of stored data is usually a storage unit, usually expressed in kb or mb, and the storage time is the cumulative time.

[0034] When either the amount of stored data or the storage time of the data packet to be analyzed is greater than or equal to the corresponding preset limit, the data packet to be analyzed will be decompressed. When there are few courses, the storage time meets the requirements, and subsequent quantification can also be carried out, which is suitable for students with a small number of students in the corresponding major.

[0035] Then extract the number of courses selected and remove courses whose number of selections is less than the preset number. Mark the number of courses selected after removal as the number of courses selected Qi. The value range of i is 1, 2, 3, ... n, where n is a positive integer, and Q1, Q2, ... Qn correspond to the number of courses selected.

[0036] The average number of selections for the remaining courses and the standard deviation of selections for the remaining courses are calculated using the formula. The preference dispersion value is obtained by the ratio of the standard deviation of selections to the average number of selections. The preference dispersion value is then multiplied by the corresponding number of selections for the remaining courses to obtain the median value for each course. Finally, several real-time median values ​​within the most recent preset time period are extracted and averaged to obtain the dynamic preference value.

[0037] Next, extract the dynamic preference values ​​of courses under the same major, sort the dynamic preference values ​​from largest to smallest, and then sort the corresponding courses by the dynamic preference values ​​to generate the first course selection recommendation set;

[0038] Obtain multiple priority course selection recommendation sets under the department, extract the weighting parameter one corresponding to the sorting, multiply the weighting parameter one by the preference dynamic value corresponding to the sorting, and then average the results of multiple multiplications to generate multiple first preference mean values. Sort the first preference mean values, and then form a second course selection recommendation set by combining the sorted first preference mean values ​​with the corresponding courses.

[0039] The system retrieves the complete set of priority course recommendations, extracts the second weighting parameter corresponding to each ranking, multiplies the second weighting parameter with the corresponding dynamic preference value, averages the results of multiple multiplications to generate multiple second preference mean values, sorts these second preference mean values, and uses the sorted second preference mean values ​​combined with the corresponding courses to form a third course recommendation set. The first, second, and third course recommendation sets are then sent to a visualization window for real-time refresh and display, enabling visualization of the recommended courses. The weighting parameter is a constant value corresponding to the ranking position; the higher the ranking, the larger the weighting parameter, and weighting parameters of the same type are subtracted by 1.

[0040] In summary, this invention stores student information through student registration, marking grade and major within the student information as matching features. It also collects student course selection information and combines this information with the matching features to generate historical tags. Analyzing a large amount of historical tags generates dynamic preference values. These dynamic preference values ​​are used to construct a first, second, and third course recommendation set. The top three courses in the first recommendation set, ranked by dynamic preference values, are then displayed prominently on the page. When a student registers and logs in, the system automatically matches them with students in the same major, recommending courses based on their major preferences, thus enhancing the correlation between students' preferences within the same major. Furthermore, the matching system of the second and third recommendation sets, along with recommendations based on the current grade level, correlates the preferences of students in the same major but not their major. This combined major and student preferences in the recommendations, making the process more professional and intelligent. On one hand, it intelligently matches and recommends courses to students in the same major and grade; on the other hand, it synchronously recommends courses to students outside their major.

[0041] Example 2:

[0042] like Figure 2 As shown, the course selection recommendation platform has a learning acquisition unit and an integration feedback unit connected by signals, and the feedback monitoring unit is connected to a division and adjustment unit.

[0043] During the course selection and learning process by students through the visual window, the learning acquisition unit is used to collect visual learning images and learning information of students during the course learning process within a preset time, and send the visual learning images and learning information of students during the course learning process to the feedback monitoring unit; the learning information includes students' course completion scores, students' course completion time, etc.; the students' course completion scores are 0-100, and the students' course completion scores of 0-100 represent their completion ability after selecting courses in this major;

[0044] The integrated feedback unit receives visual learning images within a preset time period, extracts the student's head contour from the visual learning images, marks the central axis of the visualization window as the focusing line, marks the midpoint of the focusing line as the focal point, uses the midpoint between the two eyes in the head contour as the dynamic point, constructs a line perpendicular to the focusing line, and sets the perpendicular line at the focal point. Based on the perpendicular line, draw parallel lines to the dynamic point, and extend them from the focal point to intersect the parallel lines perpendicularly to form an intersection point. Connect the focal point, dynamic point, and intersection point end to end to form a virtual perspective triangle. Collect the angle between the line connecting the focal point and dynamic point in the virtual perspective triangle and the parallel line, and record the absolute value of the angle as the virtual perspective. Then, obtain the distance between the focal point and dynamic point and mark it as the virtual distance. Multiply the virtual perspective and virtual distance with the weight parameter, and add the results to obtain the visual feedback factor. When the visual feedback factor is greater than the preset feedback value, a learning reminder signal is generated.

[0045] The generated visual feedback factor is also stored. When the visual feedback factor is less than the preset feedback value, no learning reminder signal is generated. The generation of the learning reminder signal is determined by the visual feedback factor. After the learning reminder signal is generated, the learning reminder text is edited and sent to the visualization window interface for scrolling at the top. The smaller the visual feedback factor, the more focused the student's attention is.

[0046] The average feedback value obtained by averaging the visual feedback factors throughout the entire course learning cycle is denoted as Q.

[0047] Extract student course completion scores from the learning information, divide students of the same major equally and mark them as coordinates (i, n), and then construct a student course completion score matrix and a student course completion duration matrix using the above information, as follows:

[0048]

[0049] In the above matrix R ij In this context, i represents the i-th student, j represents the j-th course, and R represents the student's completion score. The value of R ranges from [0, 100]. The higher the score, the higher the ability, and vice versa.

[0050] Then through formula In the above formula, sim(U a U i R represents the matching degree between the student to be analyzed and the students in the matrix, y represents the courses they have studied together, and R a，y and R i，y Let a and i represent the completion scores of students in course y, respectively. a and iThese represent the average scores of students completing the course, which means that the matching degree is used to judge the similarity of students' abilities in completing the course. The higher the matching degree, the better the course selection will match the ability to complete the course.

[0051] Furthermore, the matching degree of different courses in this major is extracted and averaged to obtain the mean matching value, which is then labeled as W, and normalized using the following formula: The ability correction value A of the course is obtained; e1, e2, e3 and e4 are all correction factors, which make the simulation results closer to the true value; where e1+e2+e3+e4=9.12, and e2>e4>e3>e1;

[0052] Next, the course's ability correction value and its corresponding dynamic preference value are multiplied by the second weight parameter, and the results of the multiplication are added together to obtain a new dynamic preference value. The courses in the course selection recommendation set are then sorted from largest to smallest based on the new dynamic preference value and the order of the corresponding courses is refreshed.

[0053] In summary, this invention further collects visual learning images and learning information from students during a preset learning period. It analyzes and processes these images to generate visual feedback factors, which in turn generate learning reminder signals. Once generated, the reminder text is edited and displayed on a visual window for scrolling, thus reminding students to focus. Furthermore, it normalizes and combines the learning information and feedback averages to generate a course ability correction value. This value is then combined with a corresponding dynamic preference value to generate a new dynamic preference value. The new dynamic preference value is used to sort the courses from largest to smallest, refreshing the order of courses within the recommended course selection set, further refining the course recommendations. This approach not only reminds students to focus during learning but also quantifies student learning abilities with their preferred courses, improving the accuracy of course recommendations and preventing students from blindly selecting courses, thereby enhancing the system's intelligence.

[0054] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A college course selection recommendation system based on artificial intelligence, characterized in that, It includes a visual window and a course selection recommendation platform. The visual window and the course selection recommendation platform are connected via the Internet. The visual window includes a registration and login unit and a data collection and transmission unit. The registration and login unit is used by students to register by submitting their information through a visual window. The unit marks the successful registration time and sends the registered student information to the course selection recommendation platform for storage. The student information includes the student's name, age, ID number, mobile phone number, grade, department, and major. The grade and major are marked as matching feature information. The collection and sending unit is used to collect students' course selection information and send it to the course selection recommendation platform for storage. The course selection recommendation platform is used to receive students' course selection information and match it with student information to generate historical tag information. Historical tag information with the same matching feature information is stored in the same area and packaged to generate corresponding data packets to be analyzed. The data in the data packets to be analyzed is analyzed and judged to generate dynamic preference values. The course selection recommendation set is constructed through the analysis of dynamic preference values. The course selection recommendation set is also sent to the visualization window to refresh and display the corresponding courses in real time. The specific steps for generating preference dynamic value analysis are as follows: The storage volume and / or storage time of the data packet to be analyzed are obtained. When either the storage volume or the storage time of the data packet to be analyzed is greater than or equal to the corresponding preset limit value, the data packet to be analyzed is decompressed, the number of courses selected is extracted, and courses with a selection volume less than the preset number are removed. The selection volume of the removed courses is marked as the remaining selection volume Qi. The value range of i is 1, 2, 3, ... n, where n is a positive integer, and Q1, Q2, ... Qn correspond to the remaining selection volume of courses, respectively. The mean number of choices for the remaining courses and the standard deviation of choices for the remaining courses are calculated using the formula. Then, the preference dispersion value is obtained based on the ratio of the standard deviation of choices to the mean number of choices. The preference dispersion value is then multiplied by the number of remaining courses selected to obtain the median value for each course. Several real-time median values ​​within the most recent preset time period are then extracted and averaged to obtain the dynamic preference value.

2. The college course selection recommendation system based on artificial intelligence according to claim 1, characterized in that, The analysis and construction process of the course selection recommendation set includes: Extract the dynamic preference values ​​of courses within the same major, then sort the dynamic preference values ​​from largest to smallest, and use the dynamic preference values ​​to rank the corresponding courses and generate the first set of recommended courses. Obtain multiple priority course selection recommendation sets under the department, extract the weighting parameter one corresponding to the sorting, multiply the weighting parameter one by the preference dynamic value corresponding to the sorting, and then average the results of multiple multiplications to generate multiple first preference mean values. Sort the first preference mean values, and then form a second course selection recommendation set by combining the sorted first preference mean values ​​with the corresponding courses. Obtain the complete set of priority course recommendations, extract the weighting parameter 2 corresponding to each sort, multiply the weighting parameter 2 by the corresponding dynamic preference value, average the results of multiple multiplications to generate multiple second preference mean values, sort the second preference mean values, and then form a third course recommendation set by combining the sorted second preference mean values ​​with the corresponding courses.

3. The college course selection recommendation system based on artificial intelligence according to claim 1, characterized in that, The course selection recommendation platform has a learning acquisition unit and an integration feedback unit connected by signals, and the feedback monitoring unit has a division and adjustment unit connected by signals. During the process of students selecting and learning courses through a visual window, the learning acquisition unit is used to collect visual learning images and learning information of students during the course learning process within a preset time, and sends the visual learning images and learning information of students during the course learning process to the feedback monitoring unit. The integrated feedback unit is used to receive visual learning images within a preset time, analyze and process them to generate visual feedback factors, determine the generation of learning reminder signals based on the visual feedback factors, edit the learning reminder text after the learning reminder signal is generated, and send the learning reminder text to the visualization window interface for top scrolling. It also obtains the visual feedback factor during the entire course learning cycle and averages it to obtain the feedback mean. Then, it normalizes the data in the learning information and the feedback mean and combines them to generate the course's ability correction value. Then, it multiplies the course's ability correction value and its corresponding preference dynamic value with the weight parameter two, and adds the results of the multiplication to obtain a new preference dynamic value. The new preference dynamic value is then used to sort the courses from largest to smallest and refresh the order of the corresponding courses in the course selection recommendation set.

4. The college course selection recommendation system based on artificial intelligence according to claim 3, characterized in that, The specific generation process of the visual feedback factor is as follows: Extract the student's head contour from the visual learning image. Mark the central axis of the visualization window as the focusing line and the midpoint of the focusing line as the focal point. Using the midpoint between the two eyes in the head contour as the dynamic point, construct a line perpendicular to the focusing line, with the perpendicular line placed at the focal point. Based on the perpendicular line, draw parallel lines to the dynamic point, extending from the focal point to intersect the parallel lines perpendicularly to form an intersection point. Connect the focal point, dynamic point, and intersection point end-to-end to form a virtual perspective triangle. Collect the angle between the line connecting the focal point and dynamic point in the virtual perspective triangle and the parallel line, and record the absolute value of the angle as the virtual perspective. Obtain the distance between the focal point and dynamic point and mark it as the virtual distance. Multiply the virtual perspective, virtual distance, and weight parameter one, and add the results to obtain the visual feedback factor. When the visual feedback factor is greater than the preset feedback value, a learning reminder signal is generated and stored. When the visual feedback factor is less than the preset feedback value, no learning reminder signal is generated.

5. The college course selection recommendation system based on artificial intelligence according to claim 3, characterized in that, The specific process of normalizing the capability correction values ​​is as follows: The visual feedback factors throughout the entire course learning cycle are obtained and averaged to obtain the feedback mean; then, a course completion score matrix is ​​constructed using learning information; the matching degree between the students to be analyzed and the students in the matrix is ​​obtained, and the matching degree of different courses in this major is extracted and averaged to obtain the matching mean; the feedback mean and the matching mean are normalized to obtain the course ability correction value.

Images