WO2015043073A1 - 一种关键知识点推荐方法及其系统 - Google Patents
一种关键知识点推荐方法及其系统 Download PDFInfo
- Publication number
- WO2015043073A1 WO2015043073A1 PCT/CN2013/088748 CN2013088748W WO2015043073A1 WO 2015043073 A1 WO2015043073 A1 WO 2015043073A1 CN 2013088748 W CN2013088748 W CN 2013088748W WO 2015043073 A1 WO2015043073 A1 WO 2015043073A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- knowledge point
- knowledge
- key
- relationship
- relationship strength
- Prior art date
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/20—Information retrieval; Database structures therefor; File system structures therefor of structured data, e.g. relational data
- G06F16/24—Querying
- G06F16/245—Query processing
- G06F16/2457—Query processing with adaptation to user needs
- G06F16/24578—Query processing with adaptation to user needs using ranking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N5/00—Computing arrangements using knowledge-based models
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/31—Indexing; Data structures therefor; Storage structures
- G06F16/313—Selection or weighting of terms for indexing
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/951—Indexing; Web crawling techniques
Definitions
- the invention relates to a key knowledge point recommendation method and a system thereof, and belongs to an electric digital data processing technology. Background technique
- Key knowledge points are some of the knowledge points that are relatively strongly related to other knowledge points in the field or related fields.
- Each field or related field usually has some key knowledge points. Through the study of key knowledge points, it can well understand the whole picture of the domain or related fields, thus promoting the learning of other knowledge points. Because of the text information of each field The resources are very large. With the development of the e-commerce era, a large amount of key knowledge is stored in e-books, periodicals, digital newspapers and periodicals, so when users learn a certain knowledge, quickly and timely locate and accurately track key knowledge points, it becomes An important issue.
- the key knowledge point recommendation method adopted by the prior art is based on fuzzy logic information recommendation technology, using fuzzy sets to store user files, and using fuzzy logic reasoning to match user files with received information, and filtering out information that meets user preferences. And sorted by the degree of preference to the user.
- the information system established by fuzzy logic has low calculation accuracy and poor adaptability.
- the method needs to have user files. It takes a certain time for the information collection and screening of users' preferences, and the information of users' interest is constantly Change, it is difficult to track and update user files in real time; in addition, this method can only get information or knowledge points of interest to users, but can not calculate key knowledge points.
- the key knowledge point recommendation method adopted by the prior art is an information recommendation technique based on fuzzy logic, and the fuzzy logic algorithm has low calculation precision and poor adaptability; To have a user profile, it is difficult to track and update user profiles in real time; only information or knowledge points of interest to the user can be obtained, and key knowledge points cannot be obtained. Therefore, a method and a system for recommending the key knowledge of the user according to the strength of the knowledge point based on the strength of the knowledge point, and helping the user to more effectively and selectively understand the key knowledge points are provided.
- a key knowledge point recommendation method includes the following steps: calculating a knowledge point relationship strength in a knowledge point set; according to the knowledge point relationship in the knowledge point set The strength calculates the weight of each knowledge point, and saves the knowledge point and its weight correspondingly; determines a key knowledge point according to the weight of the knowledge point, and recommends the key knowledge point to the user.
- the process of calculating the intensity of the knowledge point relationship in the knowledge point set includes: calculating the dominant relationship strength of the knowledge point; calculating the implicit relationship strength according to the intensity of the dominant point of the knowledge point; and the intensity and recessiveness according to the dominant relationship of the knowledge point Relationship strength calculates the intensity of knowledge point relationships.
- the process of calculating the intensity of the knowledge point relationship in the knowledge point set includes the following steps: calculating the dominant relationship strength of the knowledge points in the knowledge point set, and generating a knowledge point relationship strength matrix ⁇ ; constructing the intensity matrix according to the knowledge point relationship Weighted directed graph G; Calculate the implicit relationship strength of the knowledge point according to the weighted directed graph G, and generate the implicit point relationship strength matrix I of the knowledge point; traverse the knowledge point implicit relationship intensity matrix I, update the knowledge point relationship strength matrix ⁇ .
- the knowledge points in the knowledge point set and their weights are corresponding to ⁇ in the list L.
- calculating a weight of each knowledge point according to the intensity of the knowledge point relationship in the knowledge point set, and processing the knowledge point and its weight corresponding to the processing includes: obtaining, according to the knowledge point relationship strength matrix in the knowledge point set, Knowledge point. , for the collection of all sides of the starting point ⁇ and . , is the set of all edges of the end point; "; set each knowledge point., the weight is ⁇ , for the weight of each edge in ⁇ multiplied by ⁇ ⁇ and then join the weight of the knowledge point ⁇ ; for ⁇ ;"
- the weight of each edge in the multiplication is multiplied by ⁇ , and the weights obtained after the calculation is completed are in the list L; where f, is the control t, each knowledge point.
- the initial value of the weight w is set to zero.
- determining a key knowledge point according to the weight of the knowledge point, and recommending the key knowledge point to the user includes: sorting each knowledge point according to the weight of the knowledge point in descending order, and updating the list L of the knowledge points; The first K knowledge points in the list L of knowledge points in the knowledge point set are recommended as key knowledge points to the user, where ⁇ is an integer greater than or equal to 2.
- the processing of calculating the explicit relationship strength of the knowledge points in the knowledge point set includes: calculating a forward dominant dominant relationship strength of the knowledge points in the knowledge point set; and calculating a knowledge point reverse dominant relationship strength in the knowledge point set;
- the intensity of the dominant point of the knowledge point is calculated according to the strength of the positive point of the knowledge point and the intensity of the inverse dominant relationship in the knowledge point set; and the intensity of the knowledge point relationship is generated according to the dominant relationship strength of the knowledge points in the knowledge point set ⁇ .
- a key knowledge point recommendation system comprising: a relationship strength calculation module, configured to calculate a knowledge point relationship strength in a knowledge point set; a weight calculation ⁇ module, configured to calculate each knowledge point according to the knowledge point relationship strength in the knowledge point set The weight, and the knowledge points and their weights are correspondingly saved; the recommendation module is used to determine the key knowledge points according to the weight of the knowledge points, and recommend the key knowledge points to the user.
- the relationship strength calculation module is configured to calculate the intensity of the dominant relationship of the knowledge point; calculate the intensity of the implicit relationship according to the intensity of the explicit relationship of the knowledge point; and calculate the intensity of the knowledge point relationship according to the intensity of the explicit relationship of the knowledge point and the intensity of the implicit relationship.
- the explicit relationship strength calculation module specifically includes: a relationship strength matrix generating unit, configured to calculate a dominant relationship strength of the knowledge points in the knowledge point set, and generate a knowledge point relationship strength matrix ⁇ ; a weighted directed graph construction unit , configured to construct G according to the knowledge point relationship strength matrix; a recessive relationship strength matrix generating unit, configured to calculate the implicit relationship strength of the knowledge point according to the weighted directed graph G, and generate a knowledge point implicit relationship strength matrix I; , used to traverse the knowledge point implicit relationship strength matrix I, update the knowledge point relationship strength matrix ⁇ . Correspond to the knowledge points and their weights in the knowledge point set ⁇ in the list L.
- the weight calculation saving module is configured to use the knowledge point relationship strength matrix in the knowledge point set M, obtain the set f of all edges starting from the knowledge point and the set of all edges that are the end points; "; set each knowledge point., the weight is W, and multiply the weight of each edge in T by After ⁇ is added to the weight of the knowledge point; for ⁇ ; the weight of each edge in the multiplication is added to W later, and the weight obtained after the calculation is completed is in the list L; where, is the control t, each knowledge point. ,the weight of
- the recommendation module includes: a sorting sub-module, configured to arrange each knowledge point in descending order of weights of the knowledge points, and update a list L of knowledge points; a recommendation sub-module, configured to take a list of knowledge points in the knowledge point set.
- the first K knowledge points are recommended as key knowledge points to the user, where K is an integer greater than or equal to 2.
- the relationship strength matrix generating unit includes: a forward dominant relationship strength calculating subunit, configured to calculate a forward dominant dominant relationship strength of the knowledge points in the knowledge point set; and a reverse explicit relation strength calculating subunit, configured to Calculating the intensity of the inverse dominant relationship of the knowledge points in the knowledge point set; the explicit relationship strength calculation sub-unit is used to calculate the knowledge point display according to the intensity of the positive dominant relationship and the intensity of the inverse dominant relationship in the knowledge point set sexual relationship strength; The relationship strength matrix generation sub-unit is configured to generate a knowledge point relationship strength matrix M according to the dominant relationship strength of the knowledge points in the knowledge point set.
- the updating unit includes: a searching subunit, configured to traverse each item in the implicit relationship strength matrix I; a determining subunit for determining the size of Iij and My; and updating the subunit, if My for re-assignment, Mfly, the updated knowledge intensity matrix M relationship, processing in an implicit relational matrix I is the intensity; if ⁇ 3 ⁇ 4, directly next process implicit relational matrix I is the intensity, until Traversing the implicit relationship strength matrix I.
- One or more computer-readable interfaces having computer-executable instructions that, when executed by a computer, perform a critical knowledge point recommendation method, the method comprising: calculating a knowledge point relationship strength in a set of knowledge points; The knowledge point relationship strength in the calculation calculates the weight of each knowledge point, and saves the knowledge points and their weights accordingly; determines the key knowledge points according to the weight of the knowledge points, and recommends the key knowledge points to the users.
- the above technical solution has one or more of the following advantages compared with the prior art: (1)
- the key knowledge point recommendation method of the present disclosure by calculating the knowledge point relationship in the knowledge point set The intensity gets the intensity of the knowledge points, and the user learns the knowledge according to the intensity of the knowledge points to help the user learn some key knowledge points more objectively, effectively and selectively. It effectively avoids the recommendation of knowledge points in the prior art based on fuzzy logic information recommendation technology.
- the fuzzy logic algorithm has low calculation accuracy and poor adaptability; and needs to have user files, and it is difficult to track and update user files in real time; The information or knowledge points cannot be answered by key knowledge points.
- the weight of the knowledge point is based on the knowledge point relationship intensity matrix M in the knowledge point. , for all sides of the starting point and to. For the weight of all the edges of the end point, the weight of the knowledge point is calculated by the weight of the two-way edge, and the balanced treatment of the weight of the knowledge point is considered, and the accuracy of the weight of the knowledge point is improved.
- the intensity of the explicit relationship of the knowledge point is obtained by calculating the strength of the positive dominant relationship and the strength of the inverse dominant relationship, and the two-way relationship strength evaluation method further improves the explicit relationship.
- the accuracy of the intensity is obtained by calculating the strength of the positive dominant relationship and the strength of the inverse dominant relationship, and the two-way relationship strength evaluation method further improves the explicit relationship.
- the key knowledge point recommendation method of the present disclosure by converting the explicit relationship matrix into a weighted directed graph, facilitates the calculation of the shortest path between the knowledge points, and also facilitates the implementation of the algorithm, thereby improving the computational efficiency.
- Open the key knowledge point recommendation method respectively store the explicit relationship strength and the implicit relationship strength in the explicit relationship intensity matrix and the implicit relationship intensity matrix, which are convenient for taking values when performing operations, and further improve the operation. speed.
- the shortest simple path length calculation method adopts the Dijkstra algorithm, the calculation speed is fast, the fast search is realized, and the response speed is improved.
- the key knowledge point recommendation method of the present disclosure adopts SPFA Algorithm, which maintains a queue, initially adding source knowledge points to the queue. Each time a knowledge point is taken from the queue and all points adjacent to it are slackened. If an adjacent point is slack, it is enqueued. The algorithm ends until the queue is empty.
- the algorithm is simple to implement, fast in operation rate, and improves response speed.
- the shortest simple path length calculation method uses the Floyd-Warshall algorithm, can calculate the shortest path between any two points, and can usually be used in any picture, including directed graph, with negative power The edge of the graph, by considering the shortest child ⁇ to get the shortest.
- the algorithm is easy to implement, the calculation speed is fast, and the response speed is improved.
- the key knowledge point recommendation method of the present disclosure the shortest simple path length calculation method adopts the Bellman-Ford algorithm, is suitable for single source shortest path calculation, and is easy to program and easy to implement.
- the key knowledge point recommendation method of the present disclosure sets a control factor when calculating the strength of the positive dominant relationship, and effectively controls the change of the magnitude of the dominant relationship by setting the control factor, and selects according to the characteristics of the knowledge point set.
- the size of the control factor, and the control factor is optimized according to the characteristics of the knowledge point. In general, the control factor is set to 1 good effect.
- the key knowledge point recommendation method of the present disclosure sets the correlation factor a when calculating the inverse dominant relationship strength, and effectively controls the influence of the positive dominant relationship on the inverse dominant relationship by setting the correlation factor, 1 ⁇ « ⁇ 5, select different values according to the characteristics of the knowledge point set.
- the correlation factor is set to 2 to get better results.
- control parameter f, ⁇ '" is set in the weight of the calculation knowledge point, and the control parameter f is used to respectively use the edge of the knowledge point as the starting point and the edge of the end point. Controlled to adjust the contribution of the two types of edges to the weights.
- the values of the control parameters f the difference is large, the difference of the contribution of the edges is large; when the difference between the values of the control parameters ⁇ , is small, the contribution of the edges The difference is small.
- the key knowledge point recommendation method of the present disclosure taking the top K knowledge points in the knowledge point list L As a key knowledge point, it is recommended to the user.
- the value range of K is generally 2 ⁇ K ⁇ 45. Different thresholds can be selected according to the characteristics of the knowledge point set. It can be adjusted according to the user's needs, and the user is at the recommended key knowledge points. You can choose your favorite knowledge points.
- the key knowledge point recommendation system of the present disclosure effectively avoids the recommendation of knowledge points based on fuzzy logic in the prior art by using the key knowledge point recommendation method, and the fuzzy logic algorithm has low calculation precision and poor adaptability; It is necessary to have user files, and it is difficult to track and update user files in real time; only information or knowledge points of interest to users can be obtained, and key knowledge points cannot be obtained.
- FIG. 1 is a flow chart of an embodiment of a key knowledge point recommendation method of the present invention
- Figure 2 is a schematic illustration of an embodiment of a weighted directed graph of the present invention
- FIG. 3 is a structural diagram of an embodiment of a key knowledge point recommendation system of the present invention.
- FIG. 4 is a structural diagram of an embodiment of a relationship strength calculation module of the present invention.
- DETAILED DESCRIPTION OF THE EMBODIMENTS Embodiment 1 This embodiment provides a key knowledge point recommendation method.
- the flowchart is as shown in FIG. 1.
- the knowledge point in this paper is a knowledge interaction unit, which represents a concept or an entity, such as Qin Shihuang. The Tang Dynasty and the Reform Movement of 1898.
- a collection of knowledge points refers to a domain, its related fields, or a collection of partial knowledge points in a domain.
- the relevant texts of the knowledge point name and the knowledge point are shown in Table 1.
- the names are respectively recorded as ⁇ , 8 and (. ⁇ is a RL book, the text does not Contains knowledge point names, ⁇ or C.
- the C NNN key knowledge point recommendation method includes the following steps:
- calculating the strength of the knowledge point relationship in the knowledge point set includes: first calculating the dominant relationship strength of the knowledge point; and calculating the implicit relationship strength according to the intensity of the dominant point of the knowledge point; and finally, according to the intensity of the explicit relationship of the knowledge point Implicit relationship strength calculates the intensity of knowledge point relationships.
- the process of calculating the intensity of the knowledge point relationship in the knowledge point set includes the following steps:
- (i, j) is from the knowledge point. , the intensity of the positive dominant relationship to the knowledge point, / the point of knowledge at the knowledge point.
- the control factor is set to 1.
- control factor can be set to different values such as 0.5, 0.7, 1.2, 1.5, etc., and the size of the control factor controls the change of the dominant relationship strength, the user
- the size of the control factor is selected according to the characteristics of the knowledge point set, and the control factor is optimized according to the characteristics of the knowledge point set. As shown in Table 1, the number of occurrences of knowledge point B in the relevant text of knowledge point A is 2, then the positive dominant relationship strength f P (A, B) from knowledge point A to knowledge point B is
- association factor is set to 2, and in other embodiments, the association factor can be set to different values such as 1, 1.5, 3, 4, 5, and the like.
- Correlation factor Controls the influence of positive dominant relationship on the inverse dominant relationship. The smaller the value of "the smaller the value, the greater the influence of the positive relationship on the inverse relationship. The larger the value of ⁇ , the positive relationship versus the inverse relationship. The smaller the impact.
- E y is equal to zero.
- the dominant relationship strength of the knowledge point to itself is set to zero. In other embodiments, the dominant relationship strength of the knowledge point to itself may be set to 1, but has no practical meaning.
- the dominant relationship strength between the knowledge point A, the knowledge point B and the knowledge point C is sequentially calculated.
- the key knowledge point recommendation method of this embodiment the intensity of the dominant point of the knowledge point is calculated by calculating the forward display The strength of the sexual relationship and the strength of the inverse dominant relationship are obtained.
- the two-way relationship strength assessment method further improves the accuracy of the explicit relationship strength.
- the key knowledge point method of this embodiment the calculation method of the dominant relationship strength and the implicit relationship strength are obtained by the exponential function and the logarithmic function respectively, and the mathematical model is established by using the exponential function and the logarithmic function property and the relationship between the two. The concept is clever, the algorithm is simple, and easy to implement.
- the knowledge point relationship strength matrix M (current dominant relationship strength) is generated according to the dominant relationship strength between the knowledge points A, B, and C in Table 1, as shown in Table 2: Table 2. Relationship strength matrix M dominant relationship Strength)
- the vertices of the weighted directed graph G are the same as the vertices of M.
- the invention converts the explicit relationship matrix into a weighted directed graph, which facilitates the calculation of the shortest path between the knowledge points, and also facilitates the implementation of the algorithm and improves the operation efficiency.
- the weighted directed graph G is represented by a matrix.
- Table 3 the weighted directed graph G is shown in Table 3: Table 3.
- the weighted directed graph G can also be represented by FIG. 2. As shown in FIG. 2, the explicit relationship between the knowledge points can be visually represented as the edge of the weighted directed graph G shown. The knowledge point is the vertex of the weighted directed graph G.
- the key knowledge point method of this embodiment the calculation method of the dominant relationship strength and the implicit relationship strength are obtained by the exponential function and the logarithmic function respectively, and the mathematical model is established by using the exponential function and the logarithmic function property and the relationship between the two.
- the concept is clever, the algorithm is simple, and easy to implement.
- the key knowledge point recommendation method in this embodiment separately stores the dominant relationship strength and the recessive relationship strength in the explicit relationship strength matrix and the implicit relationship strength matrix, respectively, and facilitates the value when performing the operation, thereby further improving the operation speed. .
- the weight of is the weight of each edge in T multiplied by ⁇ ⁇ and then added to the weight of the knowledge point; for ⁇ ; "the weight of each edge in the multiplication is later added w, the weight will be obtained after the calculation is completed In the list L; where, is the control ⁇ t, each knowledge point. , the initial value of the weight is set to zero.
- the control ⁇ t ⁇ can take different values such as 0.5, 1.5, 2.5, 3, 3.7, 5, 7, 9, etc., taking into account the equal treatment of the weights, using the control parameters ⁇ , respectively, to the knowledge points .
- the edge of the starting point and the edge ending with ⁇ ' are used to adjust the contribution of the two types of edges to the weight.
- the control parameter "' the difference between the values is large, the difference of the contribution of the edge is large; when the difference between the control parameters and the value is small, the difference of the contribution of the edge is small.
- the key knowledge point When the value of the control parameter is large, the key knowledge point The generality and versatility are strong; otherwise, when the control parameter takes a large value, the heat of the key knowledge points is higher; otherwise, the size of the control parameter ⁇ , can also be set according to the characteristics of the knowledge point set, and the artificial Intelligent intervention.
- S106 Determine a key knowledge point according to the weight of the knowledge point, and recommend the key knowledge point to the user.
- the knowledge points and their weights are corresponding to ⁇ in the list L.
- the specific processing includes:
- the first K knowledge points in the list L of knowledge points are recommended as key knowledge points to the user.
- K is an integer greater than or equal to 2, and does not exceed the number of all knowledge points.
- the value range of ⁇ is generally 2 ⁇ 45, which can be adjusted according to the user's needs according to the different thresholds of the knowledge point set. At the same time, the user can choose his favorite knowledge in the recommended key knowledge points. point.
- the first two knowledge points W B W A are obtained from the knowledge point list L. These two knowledge points are the key knowledge points.
- the processing of calculating the implicit relationship strength of the knowledge point according to the weighted directed graph G and generating the knowledge point implicit relationship strength matrix I in this embodiment is different from that of the first embodiment, and the other steps are the same as those of the first embodiment.
- Cij the shortest simple path length from the knowledge point 0 in the weighted directed graph G to the knowledge point 0j . If knowledge points. 'There is no simple convenientness to the knowledge point, then / is equal to zero; the intensity of the implicit relationship of the knowledge point to itself is set to zero; and the implicit relationship strength ⁇ ( , 7) ⁇ is the matrix form, then the implicit relationship strength of the knowledge point is generated. Matrix I.
- the shortest simple path length calculation method uses the SPFA algorithm, which maintains a queue and initially joins the source knowledge points to the queue. Each time a knowledge point is taken from the queue and all points adjacent to it are slackened. If an adjacent point is slack, it is enqueued. The algorithm ends until the queue is empty.
- the algorithm is simple to implement, fast in operation rate, and improves response speed.
- Embodiment 3 The processing of calculating the implicit relationship strength of the knowledge point according to the weighted directed graph G and generating the knowledge point implicit relationship strength matrix I in this embodiment is different from that of the first embodiment, and the other steps are the same as those in the first embodiment.
- the processing of calculating the implicit relationship strength of the knowledge point according to the weighted directed graph G and generating the knowledge point implicit relationship strength matrix I is:
- Cij represents the shortest simple path length from the knowledge point 0 in the weighted directed graph G to the knowledge point 0j . If knowledge points.
- Implicit relationship strength matrix I The shortest simple path length calculation method uses the Floyd-Warshall algorithm to calculate the shortest path between any two points. It can usually be used in any graph, including directed graphs, graphs with negative edges, and shortest by considering the shortest subpath. path. The algorithm is easy to implement, the calculation speed is fast, and the response speed is improved.
- Embodiment 4 The processing of calculating the implicit relationship strength of the knowledge point according to the weighted directed graph G and generating the knowledge point implicit relationship strength matrix I in this embodiment is different from that of the first embodiment, and the other steps are the same as those of the first embodiment.
- the processing of calculating the implicit relationship strength of the knowledge points according to the weighted directed graph G and generating the implicit relationship strength matrix I of the knowledge points is:
- the shortest simple path length calculation method adopts Bellman-Ford algorithm, which is suitable for single source shortest path calculation, easy to program and easy to implement.
- the key knowledge point recommendation method provided in this embodiment includes: calculating a knowledge point relationship strength in the knowledge point set; calculating a weight of the knowledge point according to the knowledge point relationship strength in the knowledge point set, and corresponding the knowledge point and the weight thereof Save; Identify key knowledge points based on the weight of the knowledge points, and recommend key knowledge points to the user.
- the method obtains the knowledge point by calculating the intensity of the knowledge point relationship in the knowledge point set. Degree, according to the intensity of the knowledge point, the user learns the knowledge to recommend, and helps the user to learn some key knowledge points more objectively, effectively and selectively.
- FIG. 5 is a block diagram showing an embodiment of a key knowledge point system of the present invention.
- the key knowledge point recommendation system of this embodiment includes the following modules:
- the relationship strength calculation module 31 is configured to calculate the knowledge point relationship strength in the knowledge point set.
- the relationship strength calculation module 31 calculates the dominant relationship strength of the knowledge point; calculates the implicit relationship strength according to the intensity of the dominant point of the knowledge point; and calculates the intensity of the knowledge point relationship according to the intensity of the explicit relationship of the knowledge point and the intensity of the implicit relationship .
- Figure 4 is a block diagram showing an embodiment of the relationship strength calculation ⁇ of the present invention.
- the relationship strength calculation module 31 specifically includes: a relationship strength matrix generation unit 311, configured to calculate a dominant intensity of the knowledge points in the knowledge point set, and generate a knowledge point relationship strength matrix M.
- the relationship strength matrix generating unit specifically includes a forward dominant relationship strength calculating subunit 3111 for calculating a knowledge point forward dominant relationship strength in the knowledge point set.
- the positive point of the knowledge point is:
- (i, j) is the strength of the positive dominant relationship from knowledge point A to knowledge point ⁇
- / is the knowledge point ⁇ at the knowledge point.
- the control factor /? 1.
- the inverse dominant relationship strength calculation sub-unit 3112 is configured to calculate the inverse dominant relationship strength of the knowledge points in the knowledge point set.
- the intensity of the inverse dominant relationship of knowledge points is:
- / , 7 is the intensity of the inverse dominant relationship from the knowledge point to the knowledge point, "as a correlation factor, 1 ⁇ « ⁇ 5, "is a positive integer, / P (J, 0 is the strength of the positive dominant relationship from the knowledge point to the knowledge point.
- the explicit relationship strength calculation sub-unit 3113 used for knowledge based on the knowledge point set The point-positive dominant relationship strength and the inverse dominant relationship strength are used to calculate the intensity of the dominant point of the knowledge point.
- the relationship strength matrix generation sub-unit 3114 is configured to generate a knowledge point relationship strength matrix M according to the dominant relationship strength of the knowledge points in the knowledge point set.
- the weighted directed graph construction unit 312 is configured to construct G according to the knowledge point relationship strength matrix.
- the weighted directed graph G includes edges, weights, and vertices, where the edges and weights are set as follows: When ⁇ ⁇ , G is from the knowledge point.
- the vertices of the weighted directed graph G are the same as the vertices of M.
- the weighted directed graph G is represented by a matrix.
- the implicit relationship strength matrix generating unit 313 is configured to calculate the implicit relationship strength of the knowledge point according to the weighted directed graph G, and generate the knowledge point implicit relationship strength matrix I.
- the intensity of the implicit relationship of knowledge points is: Among them, represent knowledge points. , to the knowledge point.
- the implicit relationship strength, C y represents the knowledge point 0l to the knowledge point in the weighted directed graph G.
- the shortest simple path length If the knowledge point to the knowledge point 0 ) does not exist simple ⁇ ⁇ , then equal to zero; the knowledge point to its own implicit relationship strength is set to zero; the implicit relationship strength is saved as a matrix form, then the knowledge point implicit relationship strength matrix is generated I.
- the shortest simple path length Cy calculation method is Dijkstra algorithm, SPFA algorithm,
- the updating unit 314 is configured to traverse the knowledge point implicit relationship strength matrix I, and update the knowledge point relationship strength Matrix M.
- the weight calculation and storage module 32 is configured to calculate the weight of each knowledge point according to the intensity of the knowledge point relationship in the knowledge point set, and save the knowledge point and its weight correspondingly.
- Calculating the weight of each knowledge point according to the intensity of the knowledge point relationship in the knowledge point set, and processing the knowledge point and its weight correspondingly includes: According to the knowledge point relationship strength matrix ⁇ in the knowledge point set, the knowledge point 0l is obtained The set of all edges of the starting point ⁇ and the set of all edges ending with 0l ⁇ ;"; set each knowledge point., the weight is w t , multiply the weight of each edge in T by ⁇ ⁇ and then add knowledge The weight of the point is ⁇ , in; for ⁇ ; the weight of each edge in the multiplication is multiplied by w, and the weight obtained after the calculation is completed ⁇ is in the list L; where f, is the control ⁇ t, each knowledge point .
- the initial value of the weight is set to zero.
- the knowledge points in the knowledge point set and their weights are corresponding to ⁇ in the list L.
- Recommended module 33 configured to determine a key knowledge point according to the weight of the knowledge point, and recommend the key knowledge point to the user.
- the method includes: a sorting sub-module 331, configured to sort each knowledge point according to the weight of the knowledge point, and update the list of knowledge points.
- K is an integer greater than or equal to 2
- the key knowledge point recommendation system of the present embodiment using the key knowledge point recommendation method, effectively avoids the recommendation of knowledge points in the prior art based on fuzzy logic.
- Technology, fuzzy logic algorithm has low calculation accuracy and poor adaptability; it also needs to have user files, and it is difficult to track and update user files in real time; only information or knowledge points of interest to users can be obtained, and cannot be obtained. Key knowledge points.
- embodiments of the present invention can be provided as a method, system, or computer program product.
- the present invention can take the form of an entirely hardware embodiment, an entirely software embodiment, or a combination of software and hardware.
- the present invention may employ a computer usable storage medium (including but not limited to disk storage, in one or more of which contains computer usable program code.
- the computational instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that instructions stored in the computer readable memory produce an article of manufacture including the instruction device.
- the instruction means implements the functions specified in one or more blocks of the flow or in a flow or block diagram of the flowchart.
- These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for use on a computer or other programmable device
- the executed instructions provide steps for implementing the functions specified in one or more processes and/or block diagrams of one or more blocks.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Databases & Information Systems (AREA)
- Computational Linguistics (AREA)
- Software Systems (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
Description
Claims
Priority Applications (3)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US15/025,448 US10289623B2 (en) | 2013-09-29 | 2013-12-06 | Method and system for key knowledge point recommendation |
EP13894664.5A EP3051433A4 (en) | 2013-09-29 | 2013-12-06 | Key knowledge point recommendation method and system |
JP2016518116A JP6356789B2 (ja) | 2013-09-29 | 2013-12-06 | 主要知識ポイント推奨方法及びシステム |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201310456539.3A CN104516904B (zh) | 2013-09-29 | 2013-09-29 | 一种关键知识点推荐方法及其系统 |
CN201310456539.3 | 2013-09-29 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2015043073A1 true WO2015043073A1 (zh) | 2015-04-02 |
Family
ID=52741914
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2013/088748 WO2015043073A1 (zh) | 2013-09-29 | 2013-12-06 | 一种关键知识点推荐方法及其系统 |
Country Status (5)
Country | Link |
---|---|
US (1) | US10289623B2 (zh) |
EP (1) | EP3051433A4 (zh) |
JP (1) | JP6356789B2 (zh) |
CN (1) | CN104516904B (zh) |
WO (1) | WO2015043073A1 (zh) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105306540A (zh) * | 2015-09-24 | 2016-02-03 | 华东师范大学 | 一种获得社交网络中影响力最大的前k个节点的方法 |
CN109800300A (zh) * | 2019-01-08 | 2019-05-24 | 广东小天才科技有限公司 | 一种学习内容推荐方法及系统 |
CN109829059A (zh) * | 2019-01-18 | 2019-05-31 | 平安科技(深圳)有限公司 | 推荐知识点的方法、装置、设备及存储介质 |
CN112380429A (zh) * | 2020-11-10 | 2021-02-19 | 武汉天有科技有限公司 | 一种习题推荐方法及装置 |
CN114936315A (zh) * | 2022-04-07 | 2022-08-23 | 网易有道信息技术(北京)有限公司 | 自适应推题的方法及其相关产品 |
Families Citing this family (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20160124959A1 (en) * | 2014-10-31 | 2016-05-05 | Google Inc. | System and method to recommend a bundle of items based on item/user tagging and co-install graph |
CN104765767B (zh) * | 2015-05-12 | 2019-11-08 | 上海复勤商务咨询有限公司 | 针对智能学习的知识存储系统及其知识搜索方法 |
CN106327321A (zh) * | 2016-08-22 | 2017-01-11 | 洪婷 | 一种自动生成财务数据的方法 |
CN107507468A (zh) * | 2017-08-11 | 2017-12-22 | 北京点易通科技有限公司 | 一种知识空间有向图的自适应学习方法及系统 |
CN107562854B (zh) * | 2017-08-28 | 2020-09-22 | 云南大学 | 一种定量分析党建数据的建模方法 |
CN107622048B (zh) * | 2017-09-06 | 2021-06-22 | 南京硅基智能科技有限公司 | 一种文本模式识别方法及系统 |
CN107943853A (zh) * | 2017-11-06 | 2018-04-20 | 浙江三米教育科技有限公司 | 知识节点选择测试方法及其所应用计算机设备和存储介质 |
CN110598073B (zh) * | 2018-05-25 | 2024-04-26 | 微软技术许可有限责任公司 | 基于拓扑关系图的实体网页链接的获取技术 |
CN109062955A (zh) * | 2018-06-25 | 2018-12-21 | 广东神马搜索科技有限公司 | 项目间关联关系挖掘方法、装置、计算设备及存储介质 |
CN109885772A (zh) * | 2019-02-27 | 2019-06-14 | 上海优谦智能科技有限公司 | 基于知识图谱的教育内容个性化推荐系统 |
CN112784062B (zh) * | 2019-03-15 | 2024-06-04 | 北京金山数字娱乐科技有限公司 | 一种成语知识图谱构建方法及装置 |
CN110928920B (zh) * | 2019-11-19 | 2022-08-09 | 广东交通职业技术学院 | 一种基于改进位置社交的知识推荐方法、系统及存储介质 |
CN112699308A (zh) * | 2021-01-13 | 2021-04-23 | 敖客星云(北京)科技发展有限公司 | 一种基于聚类算法的学科题目知识点推荐方法及系统 |
CN114781710B (zh) * | 2022-04-12 | 2022-12-23 | 云南师范大学 | 一种综合学习过程及题目知识点难度特征的知识追踪方法 |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101281530A (zh) * | 2008-05-20 | 2008-10-08 | 上海大学 | 基于概念衍生树的关键词层次聚类方法 |
US20130066921A1 (en) * | 2011-09-09 | 2013-03-14 | Sri International | Adaptive ontology |
CN103077247A (zh) * | 2013-01-17 | 2013-05-01 | 清华大学 | 一种社交网络中的朋友关系传递树的建立方法 |
Family Cites Families (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JPH11272684A (ja) * | 1998-03-19 | 1999-10-08 | Mitsubishi Electric Corp | 情報提供方法および装置 |
JP2000200197A (ja) * | 1999-01-06 | 2000-07-18 | Nippon Telegr & Teleph Corp <Ntt> | 知識蓄積・選択方法、知識蓄積・選択装置及び知識蓄積・選択プログラムを記録した記録媒体 |
WO2001006398A2 (en) | 1999-07-16 | 2001-01-25 | Agentarts, Inc. | Methods and system for generating automated alternative content recommendations |
JP2005242807A (ja) * | 2004-02-27 | 2005-09-08 | Yamatake Corp | 関連知識検索装置、文章ネットワーク生成装置、文章ネットワーク生成方法、およびプログラム |
JP2008123271A (ja) * | 2006-11-13 | 2008-05-29 | Mitsubishi Electric Corp | 知識ベース検索システム及び知識ベース検索方法 |
CN101067807A (zh) * | 2007-05-24 | 2007-11-07 | 上海大学 | 文本语义的可视化表示与获取方法 |
US9367823B1 (en) * | 2007-11-09 | 2016-06-14 | Skyword, Inc. | Computer method and system for ranking users in a network community of users |
JP2009288829A (ja) * | 2008-05-27 | 2009-12-10 | Sony Corp | コンテンツ表示装置およびコンテンツ表示方法 |
CN101630314B (zh) * | 2008-07-16 | 2011-12-07 | 中国科学院自动化研究所 | 一种基于领域知识的语义查询扩展方法 |
CN102622363A (zh) * | 2011-01-28 | 2012-08-01 | 鸿富锦精密工业(深圳)有限公司 | 关联词汇搜索系统及方法 |
JP5043209B2 (ja) * | 2011-03-04 | 2012-10-10 | 楽天株式会社 | 集合拡張処理装置、集合拡張処理方法、プログラム、及び、記録媒体 |
JP5691735B2 (ja) * | 2011-03-29 | 2015-04-01 | ソニー株式会社 | コンテンツ推薦装置、推薦コンテンツの検索方法、及びプログラム |
US8719248B2 (en) * | 2011-05-26 | 2014-05-06 | Verizon Patent And Licensing Inc. | Semantic-based search engine for content |
-
2013
- 2013-09-29 CN CN201310456539.3A patent/CN104516904B/zh not_active Expired - Fee Related
- 2013-12-06 EP EP13894664.5A patent/EP3051433A4/en not_active Ceased
- 2013-12-06 WO PCT/CN2013/088748 patent/WO2015043073A1/zh active Application Filing
- 2013-12-06 US US15/025,448 patent/US10289623B2/en active Active
- 2013-12-06 JP JP2016518116A patent/JP6356789B2/ja not_active Expired - Fee Related
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101281530A (zh) * | 2008-05-20 | 2008-10-08 | 上海大学 | 基于概念衍生树的关键词层次聚类方法 |
US20130066921A1 (en) * | 2011-09-09 | 2013-03-14 | Sri International | Adaptive ontology |
CN103077247A (zh) * | 2013-01-17 | 2013-05-01 | 清华大学 | 一种社交网络中的朋友关系传递树的建立方法 |
Non-Patent Citations (1)
Title |
---|
See also references of EP3051433A4 * |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN105306540A (zh) * | 2015-09-24 | 2016-02-03 | 华东师范大学 | 一种获得社交网络中影响力最大的前k个节点的方法 |
CN109800300A (zh) * | 2019-01-08 | 2019-05-24 | 广东小天才科技有限公司 | 一种学习内容推荐方法及系统 |
CN109829059A (zh) * | 2019-01-18 | 2019-05-31 | 平安科技(深圳)有限公司 | 推荐知识点的方法、装置、设备及存储介质 |
CN112380429A (zh) * | 2020-11-10 | 2021-02-19 | 武汉天有科技有限公司 | 一种习题推荐方法及装置 |
CN114936315A (zh) * | 2022-04-07 | 2022-08-23 | 网易有道信息技术(北京)有限公司 | 自适应推题的方法及其相关产品 |
Also Published As
Publication number | Publication date |
---|---|
US20160224564A1 (en) | 2016-08-04 |
JP6356789B2 (ja) | 2018-07-11 |
JP2016538615A (ja) | 2016-12-08 |
EP3051433A1 (en) | 2016-08-03 |
EP3051433A4 (en) | 2017-06-28 |
US10289623B2 (en) | 2019-05-14 |
CN104516904A (zh) | 2015-04-15 |
CN104516904B (zh) | 2018-04-03 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2015043073A1 (zh) | 一种关键知识点推荐方法及其系统 | |
US9747526B2 (en) | Using machine learning to define user controls for photo adjustments | |
US20160358042A1 (en) | Electronic Travel Album Generating Method and Computing Device | |
JP5964248B2 (ja) | 位置選択を容易にするための方法および装置 | |
JP2016526728A (ja) | 検索推奨方法及びその装置 | |
WO2015183370A1 (en) | Optimizing non-convex function for learning to rank | |
CN103631599B (zh) | 一种拍照处理方法、系统及移动终端 | |
Cao et al. | KORS: Keyword-aware optimal route search system | |
JP2018522344A (ja) | キーワード提示方法及び装置 | |
Giladi et al. | Bourgain’s discretization theorem | |
Dojchinovski et al. | Personalised graph-based selection of Web APIs | |
WO2012017786A1 (ja) | コミュニケーション支援装置、コミュニケーション支援方法、およびコンピュータ読み取り可能な記録媒体 | |
US20180075162A1 (en) | Generating modifiers for updating search queries | |
JP2014165828A (ja) | 通信端末、品質管理システムおよびコンピュータプログラム | |
WO2015062313A1 (zh) | 一种用于处理并提供评测信息的方法与设备 | |
US20220165309A1 (en) | Video Timed Anchors | |
WO2015043068A1 (zh) | 一种知识点关系强度的度量方法及其系统 | |
Barani | Subdifferentials of perturbed distance function in Riemannian manifolds | |
JP6852610B2 (ja) | 協調スケジューリング | |
CN109343776A (zh) | 一种自定义页面生成方法、装置及终端设备 | |
CN104424274B (zh) | 数据处理的方法及装置 | |
Said et al. | A 3d approach to recommender system evaluation | |
Zhang | THE PROGRESS OF THE DEVELOPMENT CONCEPT-THE PRACTICE OF THE NEW DEVELOPMENT VIEW IN CHINA'S SOCIALIST MARKET ECONOMY | |
Dai et al. | Uniform asymptotics for orthogonal polynomials with exponential weight on the positive real axis | |
Ding et al. | A predator-prey model of Holling-type II with state dependent impulsive effects |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 13894664 Country of ref document: EP Kind code of ref document: A1 |
|
ENP | Entry into the national phase |
Ref document number: 2016518116 Country of ref document: JP Kind code of ref document: A |
|
WWE | Wipo information: entry into national phase |
Ref document number: 15025448 Country of ref document: US |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
REEP | Request for entry into the european phase |
Ref document number: 2013894664 Country of ref document: EP |
|
WWE | Wipo information: entry into national phase |
Ref document number: 2013894664 Country of ref document: EP |