Course combination recommendation method

Abstract

The embodiment of the invention provides a course combination recommendation method, and belongs to the technical field of data processing, and the method specifically comprises the steps: obtaining the operation data of a T class course set needing to be combined; obtaining a combined set of T classes of course sets; setting an index set of each course in the combination set, calculating a distance between course combinations according to the index set, and then calculating a first similarity; determining a relationship between nodes according to the first similarity, and constructing a node network between course combinations; acquiring belonging communities of all combined nodes in the node network by using a community discovery algorithm; calculating a group recommendation value of each community, and obtaining an optimal course combination community; and converting the optimal course combination community into T-class course sets, calculating a comprehensive index value, obtaining a second similarity, obtaining an optimal recommended course in all the T-class course sets according to the comprehensive index value and the second similarity, and then obtaining an optimal course combination. Through the scheme of the invention, the efficiency, accuracy and adaptability of course combination recommendation are improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of data processing technology, and in particular to a course combination recommendation method. Background Art

[0002] At present, in today's digital age, online courses have become one of the preferred ways for many people to acquire knowledge and skills. Online courses are not restricted by geographical location and can be studied anytime and anywhere, providing learners with greater flexibility and convenience. Whether it is academic knowledge, professional skills or personal interests and hobbies, you can find a rich and diverse selection in online courses. Through online courses, learners can obtain a systematic and comprehensive learning experience through various forms of teaching content such as online videos, teaching materials, practical projects, etc. At the same time, online courses also provide learners with the opportunity to interact and communicate with other learners and professional mentors around the world, promoting knowledge sharing and collaborative learning.

[0003] However, in current Internet learning systems, most recommendations are based on a single course, which makes it impossible to form a systematic learning route.

[0004] It can be seen that there is an urgent need for an efficient, accurate and adaptable course combination recommendation method. Summary of the invention

[0005] In view of this, an embodiment of the present invention provides a course combination recommendation method, which at least partially solves the problems of poor recommendation efficiency, accuracy and adaptability in the prior art.

[0006] The embodiment of the present invention provides a course combination recommendation method, comprising:

[0007] Step 1, obtaining the operation data of the T-category course set to be combined, wherein the operation data includes the school where the students who have chosen the course are located and the number of students who have chosen the course, the course teaching team and the school where they are located, the course certification category, the semester in which the course is offered, and the accumulated interactions and views;

[0008] Step 2, obtain a combination set of T course sets, where each course combination in the combination set has and only has one course in each category, and the students who enroll in the T courses in each combination at least come from the same school;

[0009] Step 3, setting an indicator set for each course in the combination set and calculating the distance between course combinations based on the indicator set to calculate the first similarity, wherein the indicator set includes a certification index, a cumulative course opening index, a cumulative interaction index, and a course selection index;

[0010] Step 4, taking the course combination as a node, determining the relationship between the nodes according to the first similarity, and constructing a node network between the course combinations;

[0011] Step 5: Use the community discovery algorithm to obtain the communities to which all combined nodes in the node network belong;

[0012] Step 6: Calculate the group recommendation value of each community based on the course certification index, cumulative course opening index, cumulative interaction index, and course selection index to obtain the optimal course combination community;

[0013] Step 7, convert the optimal course combination community into a T-category course set, calculate the comprehensive index value of each course in the T-category course set based on the certification index, faculty index and course selection index, obtain the second similarity between the course selection demand school and the school to which each course in the T-category course set belongs, and obtain the optimal recommended courses in all T-category course sets based on the comprehensive index value and the second similarity, and then obtain the optimal course combination.

[0014] According to a specific implementation of the embodiment of the present invention, step 1 specifically includes:

[0015] Get the T-type course set C that needs to be combined = {C t |t=1,2,...,T}, where C t Represents the t-th class set, namely C t ={c tb |b=1,2,...,B}, where c tb It represents the b-th course in the t-th course set, B represents the number of courses in this set, and each course contains the school where the students enrolled and the number of students enrolled, the course teaching team and the school where they are located, the course certification category, the semester in which the course is offered, and the accumulated interaction and browsing information.

[0016] According to a specific implementation of the embodiment of the present invention, step 2 specifically includes:

[0017] According to the course set C, we can filter out the combination set CP of T-type course sets = {Z i |i=1,2,...,I}, where I is the number of course combinations, Z i represents the i-th course combination. There is only one course of each category in the course combination Z, that is, each course combination Z i The number of courses in the course combination is T, and all courses in the course combination must be selected by at least n universities, where n ≥ 1. i Expressed as: Z i ={c it |t=1,2,...,T}.

[0018] According to a specific implementation of an embodiment of the present invention, the step of setting an indicator set for each course in the combination set includes:

[0019] The course certification index is formulated based on the schools and number of students enrolled in the courses included in the course combination, the course teaching team and the school where they are located, the course certification category, the semester in which the course is offered, and the accumulated interaction and browsing information. Cumulative course index Cumulative interaction index Course Selection Index Forming the set of indicators for each course combination in represents the value of the kth indicator of the tth course in the ith combination, where

[0020] Course Accreditation Index The expression is:

[0021]

[0022] Among them, 0≤a3<a2<a1≤1;

[0023] Cumulative course index The expression is:

[0024]

[0025] Where 0≤l4<l3<l2<l1≤1, α is the number of semesters the course is offered;

[0026] Cumulative interaction index The expression is:

[0027]

[0028] Where f1 is the average interaction value in the cumulative interaction index, f2 is the average browsing value in the cumulative interaction index, w1 and w2 are weights, w1≥w2 and w1+w2=1,

[0029]

[0030] Among them, 0≤e5<e4<e3<e2<e1≤1, 0≤g5<g4<g3<g2<g1≤1, β1 is the cumulative number of interactions of the course, and β2 is the cumulative number of views of the course;

[0031] Course Selection Index for:

[0032]

[0033] in, is the rank score of the λth university that chooses this course, is the flow score of the students who choose this course in the λth university,

[0034]

[0035] Among them, 0<j4<j3<j2<j1≤1, 0<m5<m4<m3<m2<m1≤1, and δ is the number of students who choose this course in a university.

[0036] According to a specific implementation of the embodiment of the present invention, the distance is expressed as:

[0037]

[0038] in, Indicates course combination Z i The courses in the tth course set, Indicates course combination Z j The courses in the tth course set, For Courses The distance between For Courses Certification index, cumulative course opening index, cumulative interaction index, course selection index, For Courses Certification index, cumulative course opening index, cumulative interaction index, and course selection index;

[0039] The expression of the first similarity is:

[0040] S(Z i ,Z j )=1-d(Z i ,Z j ).

[0041] According to a specific implementation of the embodiment of the present invention, the expression of the node network is:

[0042]

[0043] According to a specific implementation of the embodiment of the present invention, step 5 specifically includes:

[0044] According to the node network, Louvain community discovery algorithm is used to divide the community, and the course combination community V = {V p |p=1,2,...,P}, where P is the number of communities, V p Denote the pth community, in, represents the yth course combination in the pth community, And for the course collection in the community, there is

[0045] According to a specific implementation of the embodiment of the present invention, step 6 specifically includes:

[0046] According to the community have For each course in the community, the community V is obtained by calculating the average of the certification index, cumulative course opening index, cumulative interaction index, and course selection index of all courses. p Group recommended values:

[0047]

[0048] in, represents the kth index of the tth course in the yth course combination in the pth community, w k represents the weight of the indicator,

[0049] According to the group recommendation value calculation formula, the group recommendation value of each community is obtained and sorted to obtain the optimal course combination community.

[0050] According to a specific implementation of the embodiment of the present invention, step 7 specifically includes:

[0051] Combine the best courses into a community V θ The course combination in is converted into a set of T courses, that is, and Divide all courses belonging to the same category in the course combination in the community together to obtain a new set of T courses Among them, V θ ∈V and Represents the optimal course combination community V θ The course set belonging to the tth category in have

[0052] Calculate the comprehensive index value of each course in the course set VC in For each course The corresponding index value, ω k For each indicator’s weight,

[0053] According to the school to which the course selection user belongs and the school to which the course belongs, calculate the second similarity between the school to which the course selection demand school belongs and each course in the T-category course set

[0054]

[0055] Among them, the school level is first-tier university > second-tier university > third-tier university > others, 0<q3<q2<q1≤1;

[0056] According to the comprehensive index value of the course and the second similarity Calculate the comprehensive recommendation value for each type of course

[0057]

[0058] Among them, ξ1 and ξ2 are weights, and ξ1+ξ2=1;

[0059] According to the comprehensive recommended value Sort the courses in each course set, select the courses with the highest comprehensive recommendation value in each course set, and get the optimal course combination.

[0060] The course combination recommendation scheme in the embodiment of the present invention includes: step 1, obtaining the operation data of the T-category course set to be combined, wherein the operation data includes the school where the students who have chosen the course are located and the number of students who have chosen the course, the course teaching team and the school where the course is located, the course certification category, the semester in which the course is offered, and the accumulated interactions and browsing; step 2, obtaining the combination set of the T-category course set, wherein each course combination in the combination set has and only has one course in each category, and the students who have chosen the T courses in each combination at least include the same school; step 3, setting an indicator set for each course in the combination set and calculating the distance between the course combinations based on the indicator set and then calculating the first similarity, wherein the indicator set includes the certification index, the accumulated course offering index, the accumulated interaction index and the course selection index; step 4, based on the course combination Step 5: Use the community discovery algorithm to obtain the communities to which all the combined nodes in the node network belong; Step 6: Calculate the group recommendation value of each community according to the certification index, cumulative course opening index, cumulative interaction index and course selection index of the course to obtain the optimal course combination community; Step 7: Convert the optimal course combination community into a T-type course set, calculate the comprehensive index value of each course in the T-type course set according to the certification index, faculty index and course selection index, obtain the second similarity between the course selection demand school and the school to which each course in the T-type course set belongs, and obtain the optimal recommended courses in all T-type courses according to the comprehensive index value and the second similarity, and then obtain the optimal course combination.

[0061] The beneficial effects of the embodiments of the present invention are as follows: through the scheme of the present invention, attention is paid to the intrinsic correlation between courses, and the basic data and actual needs of the colleges and universities where courses are selected are fully considered. The course combination recommendation can tailor a learning path that meets the needs of each learner, ensuring the pertinence and effectiveness of learning. At the same time, the course combination uses a community discovery algorithm to divide the community, which helps to reveal the potential organizational structure and correlation rules between the course combinations, thereby improving the efficiency, accuracy and adaptability of the course combination recommendation. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0063] Figure 1 A flowchart of a course combination recommendation method provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0064] The embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0065] The following describes the embodiments of the present invention through specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work belong to the scope of protection of the present invention.

[0066] It should be noted that various aspects of the embodiments within the scope of the appended claims are described below. It should be apparent that the aspects described herein can be embodied in a wide variety of forms, and any specific structure and / or function described herein is merely illustrative. Based on the present invention, it should be understood by those skilled in the art that an aspect described herein can be implemented independently of any other aspect, and two or more of these aspects can be combined in various ways. For example, any number of aspects described herein can be used to implement the device and / or practice the method. In addition, other structures and / or functionalities other than one or more of the aspects described herein can be used to implement this device and / or practice this method.

[0067] It should also be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention. The drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0068] Additionally, in the following description, specific details are provided to facilitate a thorough understanding of the examples. However, it will be understood by those skilled in the art that the aspects described may be practiced without these specific details.

[0069] An embodiment of the present invention provides a course combination recommendation method, which can be applied to a network course recommendation process in an Internet scenario.

[0070] See also Figure 1 , which is a flow chart of a course combination recommendation method provided in an embodiment of the present invention.

[0071] like Figure 1 As shown, the method mainly comprises the following steps:

[0072] Step 1, obtaining the operation data of the T-category course set to be combined, wherein the operation data includes the school where the students who have chosen the course are located and the number of students who have chosen the course, the course teaching team and the school where they are located, the course certification category, the semester in which the course is offered, and the accumulated interactions and views;

[0073] Furthermore, the step 1 specifically includes:

[0074] Get the T-type course set C that needs to be combined = {C t |t=1,2,...,T}, where C t Represents the t-th class set, namely C t ={c tb |b=1,2,...,B}, where c tb It represents the b-th course in the t-th course set, B represents the number of courses in this set, and each course contains the school where the students enrolled and the number of students enrolled, the course teaching team and the school where they are located, the course certification category, the semester in which the course is offered, and the accumulated interaction and browsing information.

[0075] For example, we can select the advanced mathematics and linear algebra courses on the Zhihuishu platform as an example. First, we use a web crawler to obtain all the university shared courses C1 and C2 on the Zhihuishu platform that contain advanced mathematics and linear algebra content. C1 = advanced mathematics, and there are 86 courses in total. Each advanced mathematics course c 1i ,i∈(1,86) has a unique corresponding course ID. C2=Linear Algebra, there are 46 courses in total, each of which has a unique course ID. 2i ,i∈(1,46) has a unique corresponding course ID. Each course contains operational data such as the school and number of students enrolled in the course, the course teaching team and the school, the course certification category, the semester of the course, and the accumulated interactions and views.

[0076] Step 2, obtain a combination set of T course sets, where each course combination in the combination set has and only has one course in each category, and the students who enroll in the T courses in each combination at least come from the same school;

[0077] Based on the above embodiment, step 2 specifically includes:

[0078] According to the course set C, we can filter out the combination set CP of T-type course sets = {Z i |i=1,2,...,I}, where I is the number of course combinations, Z i represents the i-th course combination. There is only one course of each category in the course combination Z, that is, each course combination Z i The number of courses in the course combination is T, and all courses in the course combination must be selected by at least n universities, where n ≥ 1. i Expressed as: Z i ={c it |t=1,2,...,T}.

[0079] In the specific implementation, we obtained the possible combination courses between advanced mathematics and linear algebra courses, and found that 171 colleges and universities selected advanced mathematics and linear algebra on the platform, with a total of 139 course combinations, namely CP = {Z i i=1,2,...,139}. Among the 139 combinations, the top five combinations most chosen by colleges and universities are shown in Table 1.

[0080] Table 1

[0081]

[0082]

[0083] Step 3, setting an indicator set for each course in the combination set and calculating the distance between course combinations based on the indicator set to calculate the first similarity, wherein the indicator set includes a certification index, a cumulative course opening index, a cumulative interaction index, and a course selection index;

[0084] Based on the above embodiment, the step of setting the indicator set for each course in the combination set includes:

[0085] The course certification index is formulated based on the schools and number of students enrolled in the courses included in the course combination, the course teaching team and the school where they are located, the course certification category, the semester in which the course is offered, and the accumulated interaction and browsing information. Cumulative course index Cumulative interaction index Course Selection Index Forming the set of indicators for each course combination in represents the value of the kth indicator of the tth course in the ith combination, where

[0086] Course Accreditation Index The expression is:

[0087]

[0088] Among them, 0≤a3<a2<a1≤1;

[0089] Cumulative course index The expression is:

[0090]

[0091] Where 0≤l4<l3<l2<l1≤1, α is the number of semesters the course is offered;

[0092] Cumulative interaction index The expression is:

[0093]

[0094] Where f1 is the average interaction value in the cumulative interaction index, f2 is the average browsing value in the cumulative interaction index, w1 and w2 are weights, w1≥w2 and w1+w2=1,

[0095]

[0096] Among them, 0≤e5<e4<e3<e2<e1≤1, 0≤g5<g4<g3<g2<g1≤1, β1 is the cumulative number of interactions of the course, and β2 is the cumulative number of views of the course;

[0097] Course Selection Index for:

[0098]

[0099] in, is the rank score of the λth university that chooses this course, is the flow score of the students who choose this course in the λth university,

[0100]

[0101] Among them, 0<j4<j3<j2<j1≤1, 0<m5<m4<m3<m2<m1≤1, and δ is the number of students who choose this course in a university.

[0102] Furthermore, the distance is expressed as

[0103]

[0104] in, Indicates course combination Z i The courses in the tth course set, Indicates course combination Z j The courses in the tth course set, For Courses The distance between For Courses Certification index, cumulative course opening index, cumulative interaction index, course selection index, For Courses Certification index, cumulative course opening index, cumulative interaction index, and course selection index;

[0105] The expression of the first similarity is:

[0106] S(Z i ,Z j )=1-d(Z i ,Z j ).

[0107] In the specific implementation, four indicators are formulated, namely, the course certification index, the cumulative course index, the cumulative interaction index, and the course selection index, based on the schools and number of students enrolled in the courses included in the course combination, the course teaching team and the school where they are located, the course certification category, the semester of the course, and the cumulative interaction and browsing information. Specifically,

[0108] Course Accreditation Index for:

[0109]

[0110] Cumulative course index for:

[0111]

[0112] Where α is the number of semesters the course is offered.

[0113] Cumulative interaction index for:

[0114]

[0115] Where f1 is the average interaction value in the cumulative interaction index, and f2 is the average browsing value in the cumulative interaction index, where w1 = 0.6 and w2 = 0.4. For f1 and f2,

[0116]

[0117] Among them, β1 is the cumulative number of interactions of the course, and β2 is the cumulative number of views of the course.

[0118] Course Selection Index for:

[0119]

[0120] in is the rank score of the λth university that chooses this course, is the flow score of the students who choose this course in the λth university. For h1 and h2,

[0121]

[0122] where δ is the number of students in a university who choose this course.

[0123] The certification index, cumulative course opening index, cumulative interaction index, and course selection index corresponding to the five course combinations in Table 1 are shown in Table 2.

[0124] Table 2

[0125]

[0126] The distance and similarity between combinations are calculated using the following formulas:

[0127] Course Package Z i and Z j The distance is defined as follows

[0128]

[0129] The similarities between the course combinations are:

[0130] S(Z i ,Z j )=1-d(Z i ,Z j )

[0131] The similarities between the five course combinations in Table 1 are shown in Table 3.

[0132] Table 3

[0133] Course Combination Similarity <![CDATA[(c 11 ,c 21 ) and (c 12 ,c 22 )]]> 0.8688 <![CDATA[(c 11 ,c 21 ) and (c 13 ,c 23 )]]> 0.9108 <![CDATA[(c 11 ,c 21 ) and (c 14 ,c 24 )]]> 0.7425 <![CDATA[(c 11 ,c 21 ) and (c 15 ,c 25 )]]> 0.9325 <![CDATA[(c 12 ,c 22 ) and (c 13 ,c 23 )]]> 0.9187 <![CDATA[(c 12 ,c 22 ) and (c 14 ,c 24 )]]> 0.7588 <![CDATA[(c 12 ,c 22 ) and (c 15 ,c 25 )]]> 0.9163 <![CDATA[(c 13 ,c 23 ) and (c 14 ,c 24 )]]> 0.765 <![CDATA[(c 13 ,c 23 ) and (c 15 ,c 25 )]]> 0.86 <![CDATA[(c 14 ,c 24 ) and (c 15 ,c 25 )]]> 0.7475

[0134] Step 4, taking the course combination as a node, determining the relationship between the nodes according to the first similarity, and constructing a node network between the course combinations;

[0135] Furthermore, the expression of the node network is:

[0136]

[0137] In the specific implementation, for the topological relationship between the combinations, the similarity between the combinations needs to be not 0, and the distance between the course certification index, cumulative course opening index, cumulative interaction index, and course selection index of the courses in the course combination and the courses in other course combinations needs to be less than the threshold δ. When the difference between any indicator of any course in the two combinations is greater than the threshold δ, it is considered that the course combinations do not reach the similarity level. In this embodiment, the threshold δ = 0.2, and the topological relationship between the combinations is expressed as:

[0138]

[0139] The topological relationship between the five course combinations in Table 1 is shown in Table 4.

[0140] Table 4

[0141] Course Combination <![CDATA[σ(Z i ,WITH j )]]> <![CDATA[(c 11 ,c 21 ) and (c 12 ,c 22 )]]> 0 <![CDATA[(c 11 ,c 21 ) and (c 13 ,c 23 )]]> 0 <![CDATA[(c 11 ,c 21 ) and (c 14 ,c 24 )]]> 0 <![CDATA[(c 11 ,c 21 ) and (c 15 ,c 25 )]]> 0 <![CDATA[(c 12 ,c 22 ) and (c 13 ,c 23 )]]> 0 <![CDATA[(c 12 ,c 22 ) and (c 14 ,c 24 )]]> 0 <![CDATA[(c 12 ,c 22 ) and (c 15 ,c 25 )]]> 0 <![CDATA[(c 13 ,c 23 ) and (c 14 ,c 24 )]]> 0 <![CDATA[(c 13 ,c 23 ) and (c 15 ,c 25 )]]> 1 <![CDATA[(c 14 ,c 24 ) and (c 15 ,c 25 )]]> 0

[0142] Step 5: Use the community discovery algorithm to obtain the communities to which all combined nodes in the node network belong;

[0143] Based on the above embodiment, step 5 specifically includes:

[0144] According to the node network, Louvain community discovery algorithm is used to divide the community, and the course combination community V = {V p |p=1,2,...,P}, where P is the number of communities, V p Denote the pth community, in, represents the yth course combination in the pth community, And for the course collection in the community, there is

[0145] In the specific implementation, the Louvain community discovery algorithm is used to divide the community for the node network between the constructed course combinations. The Louvain community discovery algorithm is a greedy-based algorithm for detecting community structures in networks. The main goal of the algorithm is to divide the nodes in the network into different communities so that the nodes within each community have more internal connections and fewer connections with nodes in other communities. The Louvain algorithm is divided into two stages: the first stage involves initializing each course combination node as an independent community and trying to move the course combination node to the adjacent community to improve the modularity; the second stage constructs a new network by treating the community node as a single node, and then applies the optimization process of the first stage again. In the Louvain algorithm, a modularity gain criterion is defined to measure the increase in modularity after moving a node to an adjacent community. This gain criterion is usually used to decide whether the move should be made. The modularity gain formula is expressed as:

[0146]

[0147] Among them, Σ in Represents node Z i The sum of the connection weights of other nodes in its community (i.e., node Z i ) Represents node Z i The degree of the node is the sum of the connection weights of other nodes in its community; Σ tot Represents node Z i The sum of the connection weights of all nodes in the entire network (i.e., node Z i total degree); Represents node Z i degree; m represents the total weight of the connection in the entire network, that is, m is equal to σ(Z i ,Z j ) has a value of 1.

[0148] Get the course combination community V = {V p |p=1,2,...,P}, where P is the number of communities, V p Denote the pth community, denoted as in, represents the yth course combination in the pth community. And for the course collection in the community, there is

[0149] The obtained course combination community distribution is shown in Table 5.

[0150] Table 5

[0151] Community <![CDATA[V1]]> <![CDATA[V2]]> <![CDATA[V3]]> <![CDATA[V4]]> <![CDATA[V5]]> <![CDATA[V6]]> <![CDATA[V7]]> Number of members 26 1 19 27 31 36 1

[0152] Step 6: Calculate the group recommendation value of each community based on the course certification index, cumulative course opening index, cumulative interaction index, and course selection index to obtain the optimal course combination community;

[0153] Based on the above embodiment, step 6 specifically includes:

[0154] According to the community have For each course in the community, the community V is obtained by calculating the average of the certification index, cumulative course opening index, cumulative interaction index, and course selection index of all courses. p Group recommended values:

[0155]

[0156] in, represents the kth index of the tth course in the yth course combination in the pth community, w k represents the weight of the indicator,

[0157] According to the group recommendation value calculation formula, the group recommendation value of each community is obtained and sorted to obtain the optimal course combination community.

[0158] In specific implementation, the group recommendation value of each community is calculated based on the calculated course certification index, cumulative course opening index, cumulative interaction index, and course selection index.

[0159] According to the community have Community V p The formula for calculating the group recommendation value is as follows

[0160]

[0161] in, represents the kth index of the tth course in the yth course combination in the pth community; w k Represents the weight of the indicator, weight w k The role of is to make personalized adjustments based on the user's specific needs for these four indicators, so that the course combination community recommendation results are more in line with actual needs and user expectations. In this embodiment, w1=w2=w3=w4=0.25. According to the group recommendation value calculation formula, the group recommendation value of each community is obtained and sorted to obtain the optimal course combination community. The group recommendation values ​​of each community are shown in Table 6.

[0162] Table 6

[0163]

[0164] Step 7, convert the optimal course combination community into a T-category course set, calculate the comprehensive index value of each course in the T-category course set based on the certification index, faculty index and course selection index, obtain the second similarity between the course selection demand school and the school to which each course in the T-category course set belongs, and obtain the optimal recommended courses in all T-category course sets based on the comprehensive index value and the second similarity, and then obtain the optimal course combination.

[0165] Based on the above embodiment, step 7 specifically includes:

[0166] Combine the best courses into a community V θ The course combination in is converted into a set of T courses, that is, and Divide all courses belonging to the same category in the course combination in the community together to obtain a new set of T courses Among them, V θ ∈V and Represents the optimal course combination community V θ The course set belonging to the tth category in have

[0167] Calculate the comprehensive index value of each course in the course set VC in For each course The corresponding index value, ω k For each indicator’s weight,

[0168] According to the school to which the course selection user belongs and the school to which the course belongs, calculate the second similarity between the school to which the course selection demand school belongs and each course in the T-category course set

[0169]

[0170] Among them, the school level is first-tier university > second-tier university > third-tier university > others, 0<q3<q2<q1≤1;

[0171] According to the comprehensive index value of the course and the second similarity Calculate the comprehensive recommendation value for each type of course

[0172]

[0173] Among them, ξ1 and ξ2 are weights, and ξ1+ξ2=1;

[0174] According to the comprehensive recommended value Sort the courses in each course set, select the courses with the highest comprehensive recommendation value in each course set, and get the optimal course combination.

[0175] In specific implementation, according to the optimal course combination community V θ (Where V θ ∈V and XV θ =max{XV P}), the community V θ The course combination in is converted into a set of T courses, that is, and Divide all courses belonging to the same category in the course combination in the community together to obtain a new set of T courses in Represents the optimal course combination community V θ The course set belonging to the tth category in have Calculate the comprehensive index value of each course in the course set VC in For each course The corresponding index value, ω k For the weight of each indicator, similarly, users can make personalized adjustments according to their needs.

[0176] Specifically, the course combination in the optimal community V6 obtained in step 6 is converted into two types of courses: advanced mathematics and linear algebra, and the comprehensive index value of the courses in each type of course is calculated. At the same time, users can choose suitable courses according to their preferences. For example, if a user attaches more importance to a certain index among the course certification index, cumulative course opening index, cumulative interaction index, and course selection index, he can set ω k The value of completes the calculation of the course comprehensive index. In this embodiment, ω1=ω2=ω3=ω4=0.25. The comprehensive index of each course in the advanced mathematics course set is shown in Table 7, and the comprehensive index of each course in the linear algebra course set is shown in Table 8:

[0177] Table 7

[0178]

[0179] Table 8

[0180]

[0181] In addition, based on the school to which the course selection user belongs and the school to which the course belongs, the similarity between the course selection school and the school to which each course in the T-category course set belongs is calculated. Specifically

[0182]

[0183] Among them, the school levels are 985>211>undergraduate>others.

[0184] According to the comprehensive index value of the course and similarity Calculate the comprehensive recommendation value for each type of course The calculation formula is Where ξ1 and ξ2 are weights, and ξ1+ξ2=1. In this embodiment, ξ1=ξ2=0.5. Sort the courses in each category, select the course with the highest comprehensive recommendation value in each category, and obtain the optimal course combination. As shown in Table 9 and Table 10:

[0185] Table 9

[0186]

[0187] Table 10

[0188]

[0189] From Table 9 and Table 10, we can see that the advanced mathematics courses Linear Algebra Course The comprehensive recommendation value is the highest, which is the optimal course combination.

[0190] In addition, this embodiment also provides personalized recommendations. After obtaining the comprehensive recommendation values ​​of advanced mathematics and linear algebra, users can filter the recommendation results according to their preferences. If the platform certifications of advanced mathematics and linear algebra that user 1 wants to select are both national-level courses, the filtered courses are shown in Table 11:

[0191] Table 11

[0192]

[0193] From Table 11, we can see that the 11 courses recommended for advanced mathematics are all national courses. The courses are national courses. According to the comprehensive recommended value of each course, the best course combination is the course and courses Users can also freely combine the recommended results. If user 2, in addition to user 1, also wants the course to be offered by a 985 university or a 211 university, the selected courses are shown in Table 12:

[0194] Table 12

[0195]

[0196]

[0197] From Table 12, we can see that if user 2 needs a national-level course and the course is offered by a 985 university or a 211 university, the qualified advanced mathematics courses are and Eligible linear algebra courses are As can be seen from Table 12, the comprehensive recommended values ​​of qualified advanced mathematics courses are not very different, and users can freely choose an advanced mathematics course to combine with a linear algebra course.

[0198] The course combination recommendation method provided in this embodiment obtains the operation data of T-category course sets that need to be combined, including the schools where the students who have chosen the courses are located and the number of students who have chosen the courses, the course teaching team and the school where the course is located, the course certification category, the semester in which the course is offered, and the accumulated interactions and views; obtains a combination set of T-category course sets according to the operation data, each combination has and only has one course of each category (that is, each combination is composed of T courses), and the students who have chosen the T courses in each combination are from at least the same school; obtains the certification index, the accumulated course opening index, the accumulated interaction index, and the course selection index of each course in the combination set, and calculates the distance and similarity between the combinations based on the obtained four indexes; takes the course combination as a node, determines the relationship between the nodes based on the similarity between the combinations, and constructs a node network between the course combinations; and operates the node network between the course combinations based on the constructed node network between the course combinations. Use the community discovery algorithm to obtain the communities to which all combination nodes belong; calculate the group recommendation value of each community based on the course's certification index, cumulative course opening index, cumulative interaction index, and course selection index to obtain the optimal course combination community; finally, convert the optimal community into a specific T-category course set, and combine the comprehensive index value of the course and the similarity between the school with course selection demand and the school to which the course belongs to accurately select the optimal course combination for the school with demand, thereby achieving refined course combination recommendation; compared with the existing recommendation technology based on a single course, the course combination recommendation can tailor a learning path that meets the needs of each learner, ensuring the pertinence and effectiveness of learning. At the same time, the course combination uses the community discovery algorithm to divide the community, which helps to reveal the potential organizational structure and association rules between course combinations, and ensure the accuracy of course combination recommendations.

[0199] It should be understood that various parts of the present invention can be implemented by hardware, software, firmware or a combination thereof.

[0200] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A course combination recommendation method, characterized in that: include: Step 1, obtaining the operation data of the T-category course set to be combined, wherein the operation data includes the school where the students who have chosen the course are located and the number of students who have chosen the course, the course teaching team and the school where they are located, the course certification category, the semester in which the course is offered, and the accumulated interactions and views; Step 2, obtain a combination set of T course sets, where each course combination in the combination set has and only has one course in each category, and the students who enroll in the T courses in each combination at least come from the same school; Step 3, setting an indicator set for each course in the combination set and calculating the distance between course combinations based on the indicator set to calculate the first similarity, wherein the indicator set includes a certification index, a cumulative course opening index, a cumulative interaction index, and a course selection index; Step 4, taking the course combination as a node, determining the relationship between the nodes according to the first similarity, and constructing a node network between the course combinations; Step 5: Use the community discovery algorithm to obtain the communities to which all combined nodes in the node network belong; Step 6: Calculate the group recommendation value of each community based on the course certification index, cumulative course opening index, cumulative interaction index, and course selection index to obtain the optimal course combination community; Step 7, convert the optimal course combination community into a T-category course set, calculate the comprehensive index value of each course in the T-category course set based on the certification index, faculty index and course selection index, obtain the second similarity between the course selection demand school and the school to which each course in the T-category course set belongs, and obtain the optimal recommended courses in all T-category course sets based on the comprehensive index value and the second similarity, and then obtain the optimal course combination.

2. The method according to claim 1, characterized in that , the step 1 specifically includes: Get the T-type course set C that needs to be combined = {C t |t=1,2,...,T}, where C t Represents the t-th class set, namely C t ={c tb |b=1,2,...,B}, where c tb It represents the b-th course in the t-th course set, B represents the number of courses in this set, and each course contains the school where the students enrolled in the course and the number of students enrolled in the course, the course teaching team and the school where the course is located, the course certification category, the semester in which the course is offered, and the accumulated interaction and browsing information.

3. The method according to claim 2, characterized in that , the step 2 specifically includes: According to the course set C, we can filter out the combination set CP of T-type course sets = {Z i |i=1,2,...,I}, where I is the number of course combinations, Z i represents the i-th course combination. There is only one course of each category in the course combination Z, that is, each course combination Z i The number of courses in the course combination is T, and all courses in the course combination must be selected by at least n universities, where n ≥ 1. i Expressed as: Z i ={c it |t=1,2,...,T}.

4. The method according to claim 3, characterized in that ,The step of setting the indicator set for each course in the ,combination set includes: The course certification index is formulated based on the schools and number of students enrolled in the courses included in the course combination, the course teaching team and the school where they are located, the course certification category, the semester in which the course is offered, and the accumulated interaction and browsing information. Cumulative course index Cumulative interaction index Course Selection Index Forming the set of indicators for each course combination in represents the value of the kth indicator of the tth course in the ith combination, where Course Accreditation Index The expression is: Among them, 0≤a3<a2<a1≤1 Cumulative course index The expression is: Where 0≤l4<l3<l2<l1≤1, α is the number of semesters the course is offered; Cumulative interaction index The expression is: Where f1 is the average interaction value in the cumulative interaction index, f2 is the average browsing value in the cumulative interaction index, w1 and w2 are weights, w1≥w2 and w1+w2=1, Among them, 0≤e5<e4<e3<e2<e1≤1, 0≤g5<g4<g3<g2<g1≤1, β1 is the cumulative number of interactions of the course, and β2 is the cumulative number of views of the course; Course Selection Index for: in, is the rank score of the λth university that chooses this course, is the flow score of the λth university that chooses this course, Among them, 0<j4<j3<j2<j1≤1, 0<m5<m4<m3<m2<m1≤1, and δ is the number of students who choose this course in a university.

5. The method according to claim 4, characterized in that , the distance expression is in, Indicates course combination Z i The courses in the tth course set, Indicates course combination Z j The courses in the tth course set, For Courses The distance between For Courses Certification index, cumulative course opening index, cumulative interaction index, course selection index, For Courses Certification index, cumulative course opening index, cumulative interaction index, and course selection index; The expression of the first similarity is: S(Z i ,Z j )=1-d(Z i ,Z j )。 6. The method according to claim 5, characterized in that , the expression of the node network is 7. The method according to claim 6, characterized in that , the step 5 specifically includes: According to the node network, Louvain community discovery algorithm is used to divide the community, and the course combination community V = {V p |p=1,2,...,P}, where P is the number of communities, V p Denote the pth community, in, represents the yth course combination in the pth community, And for the course collection in the community, there is 8. The method according to claim 7, characterized in that , the step 6 specifically includes: According to the community have For each course in the community, the community V is obtained by calculating the average of the certification index, cumulative course opening index, cumulative interaction index, and course selection index of all courses. p Group recommended values: in, represents the kth index of the tth course in the yth course combination in the pth community, w k represents the weight of the indicator, According to the group recommendation value calculation formula, the group recommendation value of each community is obtained and sorted to obtain the optimal course combination community.

9. The method according to claim 8, characterized in that , the step 7 specifically includes: Combine the best courses into a community V θ The course combination in is converted into a set of T courses, that is, and Divide all courses belonging to the same category in the course combination in the community together to obtain a new set of T courses Among them, V θ ∈V and Represents the optimal course combination community V θ The course set belonging to the tth category in have Calculate the comprehensive index value of each course in the course set VC in For each course The corresponding index value, ω k For each indicator’s weight, According to the school to which the course selection user belongs and the school to which the course belongs, calculate the second similarity between the school to which the course selection demand school belongs and each course in the T-category course set Among them, the school level is first-tier university > second-tier university > third-tier university > others, 0<q3<q2<q1≤1; According to the comprehensive index value of the course and the second similarity Calculate the comprehensive recommendation value for each type of course Among them, ξ1 and ξ2 are weights, and ξ1+ξ2=1; According to the comprehensive recommended value Sort the courses in each course set, select the courses with the highest comprehensive recommendation value in each course set, and get the optimal course combination.

Images