Search

CN-122021819-A - Reading path generation method based on borrowing data

CN122021819ACN 122021819 ACN122021819 ACN 122021819ACN-122021819-A

Abstract

The invention relates to the technical field of data analysis and processing, in particular to a reading path generation method based on borrowing data, which constructs a reader feature vector containing borrowing breadth, borrowing depth, time sequence rule, discipline preference and path following degree for each reader based on history borrowing records of readers; the method comprises the steps of clustering a user group by adopting a density clustering algorithm based on sharing nearest neighbor similarity to obtain at least two reader groups, carrying out time sequence borrowing sequence mining on borrowing records in each reader group, extracting frequent time sequence modes meeting time sequence constraint conditions, respectively constructing knowledge dependency graphs comprising book nodes and directed time sequence edges for each reader group based on the frequent time sequence modes, generating a structured reading path from a starting point set to a target book by adopting a graph searching algorithm according to the knowledge dependency graphs of the target book and the reader group, and generating the structured reading path with time sequence dependency relationships and fusion group differentiation characteristics based on borrowing data.

Inventors

  • WU SUYUN
  • LIU JINJIA

Assignees

  • 泉州工艺美术职业学院

Dates

Publication Date
20260512
Application Date
20260413

Claims (8)

  1. 1. The reading path generation method based on borrowing data is characterized by comprising the following steps: S1, constructing a reader feature vector for each reader based on historical borrowing records of readers, wherein feature dimensions of the reader feature vector comprise borrowing breadth, borrowing depth, time sequence rules, discipline preference and path following degree; S2, clustering and dividing the user groups by adopting a density clustering algorithm based on shared nearest neighbor similarity according to the reader feature vector to obtain at least two reader groups; S3, carrying out time sequence borrowing sequence mining on borrowing records in each reader group by adopting a sliding window and improvement PrefixSpan algorithm, and extracting frequent time sequence modes meeting time sequence constraint conditions; S4, based on the frequent time sequence mode, respectively constructing a knowledge dependent graph containing book nodes and directed time sequence edges for each reader group; S5, generating a structured reading path from a starting point set to the target book by adopting a graph search algorithm according to a knowledge dependent graph of the target book and the reader group, wherein the structured reading path comprises a main path and at least one front branch path.
  2. 2. The method for generating a reading path based on borrowed data as recited in claim 1, wherein the calculation of the shared nearest neighbor similarity in step S2 is expressed as: ; Wherein, the Representing readers Is a set of k-nearest neighbors of (c), Representing readers Is a set of k-nearest neighbors of (c), Representing readers Is a function of the reader's feature vector, Representing readers Is a function of the reader's feature vector, Representing readers And the reader Is shared with nearest neighbor similarity.
  3. 3. The reading path generating method based on borrowing data as defined in claim 1, wherein the step S3 comprises the steps of: In the reader group, dividing the borrowing sequence of each reader by adopting a sliding time window, and forming a reading session sequence by borrowing records in the same time window; Mining a time sequence dependency rule X-Y in the reading session sequence, wherein X-Y represents reading books Y in the time window after reading books X; and calculating the time sequence support degree, the time sequence confidence degree and the time sequence lifting degree of the time sequence dependency rule, and screening the rule meeting the preset threshold as a frequent time sequence mode.
  4. 4. A reading path generating method based on borrowing data as defined in claim 3, wherein the step S4 comprises the steps of: mapping each book into a book node, and recording the classification number, the subject word and the difficulty level of the book as node attributes; and constructing a directional time sequence edge pointing to the node Y from the node X for each frequent time sequence mode X-Y, and calculating the weight of the directional time sequence edge to obtain the knowledge dependency graph.
  5. 5. The method for generating a reading path based on borrowed data as recited in claim 4, the method is characterized in that the weight of the directional time sequence edge The calculation of (2) is expressed as: ; Wherein Conf temp represents the time series confidence, AVGINTERVAL represents the average borrowing interval days, T max represents the maximum window width, div (X, Y) represents the subject variability factor of book X and book Y, 、 And (3) with Is a preset weight coefficient.
  6. 6. The method of claim 1, wherein generating a structured reading path from the origin set to the target book using a graph search algorithm in step S5 comprises: Identifying a book set S G of the reading start point of the reader group according to the target book set T, the reader group G and the knowledge dependent graph DG G ; The minimum cost Steiner tree linking S G and T is constructed on DG G from which the main path and necessary pre-branches from the starting book set S G to the target book set T are extracted.
  7. 7. The method of claim 6, wherein the determining of the book set S G includes: And counting the occurrence frequency and the proportion of each book which is used as the first borrowing in the reader group G, and adding books with the occurrence frequency or the proportion exceeding a preset threshold value into the starting point set.
  8. 8. A borrowing data based reading path generation terminal comprising a processor, a memory and a computer program stored in the memory and executable on the processor, characterized in that the processor implements the steps of a borrowing data based reading path generation method as claimed in any one of the preceding claims 1-7 when executing the computer program.

Description

Reading path generation method based on borrowing data Technical Field The invention relates to the technical field of data analysis and processing, in particular to a reading path generation method based on borrowed data. Background At present, the library borrowing data mining is mainly applied to the following directions: (1) And the association rule recommendation is that the co-borrowing relation among books is mined based on the Apriori algorithm and the like, and a recommendation rule that a user borrowing A borrows B is generated. (2) And intelligent index recommendation, namely calculating an adaptation evaluation value to recommend books based on the historical borrowing times, borrowing time and other behavior characteristics of readers. (3) And generating a candidate browsing path based on a data model (such as UML and XML), and calculating the importance of the path through entropy values for screening. (4) And generating a learning path, namely constructing the association relation among documents based on the similarity of the knowledge elements to form the knowledge learning path. Based on the above application, the existing scheme has the following technical defects: (1) Timing is lost. The association rule only reflects the co-occurrence relationship, and the time sequence dependence of 'read-first A and read-second B' cannot be reflected. (2) And (5) group fracturing. The differentiated path pattern of the different background readers is not distinguished. (3) The pre-knowledge is ignored. The pre-reading basis required for understanding the target book cannot be identified. (4) The path structure is single. The output is an unordered list or a simple sequence, and the graph structure of branch convergence is lacked. Disclosure of Invention The invention aims to solve the technical problem of providing a reading path generation method based on borrowing data, which can generate a reading path with a time sequence dependency relationship, fusion group differentiation characteristics and a prepositioned knowledge structure based on the borrowing data. In order to solve the technical problems, the invention adopts the following technical scheme: a reading path generation method based on borrowing data comprises the following steps: S1, constructing a reader feature vector for each reader based on historical borrowing records of readers, wherein feature dimensions of the reader feature vector comprise borrowing breadth, borrowing depth, time sequence rules, discipline preference and path following degree; S2, clustering and dividing the user groups by adopting a density clustering algorithm based on shared nearest neighbor similarity according to the reader feature vector to obtain at least two reader groups; S3, carrying out time sequence borrowing sequence mining on borrowing records in each reader group by adopting a sliding window and improvement PrefixSpan algorithm, and extracting frequent time sequence modes meeting time sequence constraint conditions; S4, based on the frequent time sequence mode, respectively constructing a knowledge dependent graph containing book nodes and directed time sequence edges for each reader group; S5, generating a structured reading path from a starting point set to the target book by adopting a graph search algorithm according to a knowledge dependent graph of the target book and the reader group, wherein the structured reading path comprises a main path and at least one front branch path. In order to solve the technical problems, the invention adopts another technical scheme that: The reading path generating terminal based on the borrowing data comprises a processor, a memory and a computer program which is stored in the memory and can run on the processor, wherein the steps in the reading path generating method based on the borrowing data are realized when the processor executes the computer program. The reading path generation method based on borrowing data has the advantages that reader groups with similar reading behavior patterns can be identified by constructing multidimensional reader feature vectors and adopting an improved density clustering algorithm to divide reader groups, a foundation is laid for subsequent differential path generation, reading sequences with real prepositions can be identified by adopting a sliding window and an improved PrefixSpan algorithm to mine a time sequence dependency pattern, time sequence dependency is reflected, different reader groups can be respectively constructed by constructing knowledge dependency graphs, differential reading paths can be generated, the problem of group splitting is solved, and the reading dependency relationship can be intuitively displayed by adopting a graph searching algorithm to generate a structured reading path comprising a main path and prepositions, so that readers can understand the needed prepositions of target books. Drawings FIG. 1 is a flowchart of a method for generating a re