CN109344293B - Knowledge association-based theme map conflict detection method and system - Google Patents
Knowledge association-based theme map conflict detection method and system Download PDFInfo
- Publication number
- CN109344293B CN109344293B CN201810918417.4A CN201810918417A CN109344293B CN 109344293 B CN109344293 B CN 109344293B CN 201810918417 A CN201810918417 A CN 201810918417A CN 109344293 B CN109344293 B CN 109344293B
- Authority
- CN
- China
- Prior art keywords
- knowledge
- knowledge elements
- correlation
- calculating
- elements
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 23
- 238000000034 method Methods 0.000 claims abstract description 37
- 238000004364 calculation method Methods 0.000 claims abstract description 28
- 239000013598 vector Substances 0.000 claims abstract description 10
- 238000004422 calculation algorithm Methods 0.000 claims description 8
- 238000010276 construction Methods 0.000 claims description 8
- 230000008520 organization Effects 0.000 claims description 8
- 230000001902 propagating effect Effects 0.000 claims description 3
- 230000000717 retained effect Effects 0.000 claims description 3
- 230000011218 segmentation Effects 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 238000013459 approach Methods 0.000 claims description 2
- 238000000605 extraction Methods 0.000 claims description 2
- 238000012545 processing Methods 0.000 claims description 2
- 238000010586 diagram Methods 0.000 description 10
- 230000000875 corresponding effect Effects 0.000 description 5
- 230000009286 beneficial effect Effects 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 230000001149 cognitive effect Effects 0.000 description 3
- 238000004590 computer program Methods 0.000 description 2
- 230000002776 aggregation Effects 0.000 description 1
- 238000004220 aggregation Methods 0.000 description 1
- 238000009960 carding Methods 0.000 description 1
- 230000002596 correlated effect Effects 0.000 description 1
- 238000013500 data storage Methods 0.000 description 1
- 230000002708 enhancing effect Effects 0.000 description 1
- 239000000835 fiber Substances 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention belongs to the teachingThe technical field of education, and discloses a topic map conflict detection method and system based on knowledge association, which comprises the following steps: calculating the correlation R between knowledge elements in the theme mapkThe method is based on the calculation of the association relation of the path and the depth knowledge elements; calculating the correlation R between resource groups (containing knowledge elements) associated with the knowledge elements in the subject graphrCalculating the similarity between the knowledge elements by combining the resource description information associated with the knowledge elements; calculation of RkConsistency of deviation from Rr, correlation R between the derived knowledge elements calculated by way of path and depthkRelevance R calculated by utilizing word vectors constructed by resource group multi-dimensional description information labelsrThe consistency of the two in the degree of deviation is calculated. The invention is helpful to enhance the knowledge structure combing and the relevance of the resources related to knowledge, and is helpful to learners to better and faster build a personalized knowledge system.
Description
Technical Field
The invention belongs to the technical field of education, and particularly relates to a theme map conflict detection method and system based on knowledge association.
Background
Currently, the current state of the art commonly used in the industry is such that:
the topic map is one of effective knowledge aggregation methods, and the biggest advantage of the topic map is that the customization of a topic builder on knowledge composition and knowledge range can be fully exerted. Therefore, the learner has stronger relevance to the knowledge structure carding and the knowledge-related resources in the learning process, and the learner is facilitated to quickly construct an individualized knowledge system. However, the unreasonable situation exists in which the learner learns the actually constructed knowledge system in the depth and the cognitive breadth, and under the situation, two situations exist, namely, the name and the identification of the tuple are defined by the constructor and may be defined deviation; secondly, the related knowledge elements under the tuple are individually related to wrong ranges, so that inconsistency may exist among all the related knowledge elements under the tuple, and the knowledge belonging ranges represented by most knowledge elements and a small part of knowledge elements may not be under the description range of the same tuple. Therefore, unreasonable problems may exist in the subject map constructed by the learner, and the scope of the knowledge field represented by the subject map is easy to be inconsistent.
In summary, the problems of the prior art are as follows:
if the mode of constructing the subject graph is different, the mode of adopting conflict detection is different, if the mode of constructing the subject graph based on the knowledge graph is adopted to construct the data structure of the subject graph, the composition mode is different, the knowledge graph is suitable for large-scale relation construction, the method is not suitable for constructing the personalized subject graph from the aspects of the number and the relation of the constructed entities, the method is suitable for large-scale semantic relation description, the description framework of the knowledge graph mainly comprises triples, and the relation schema is restricted, a complete solution of a semantic network is needed, and the construction is time-consuming and labor-consuming.
The existing theme map is limited to the fact that the learner learns the knowledge system which is actually constructed according to the learning depth and the cognitive breadth, and is unreasonable, under the situation, two situations can exist:
(1) the name and identification of the tuple are customized by the builder, and may be a defined deviation;
(2) the related knowledge elements under the tuple are respectively related to wrong ranges, so that inconsistency may exist among the related knowledge elements under the tuple, and the knowledge belonging ranges represented by most knowledge elements and a small part of knowledge elements are not in the description range of the same tuple;
therefore, unreasonable problems may exist in the subject map constructed by the learner, and the scope of the knowledge field represented by the subject map is easy to be inconsistent.
The difficulty and significance for solving the technical problems are as follows:
the problem map conflict detection method based on knowledge association provided by the invention has the difficulty that vector-based similarity needs to be constructed for each knowledge element and knowledge element group in the problem map, and expression of the knowledge element group is expanded by adopting resource group description information.
The method has the significance of rapidly identifying the reasonable relation between the knowledge group constructed by the learner and the resource in real time and providing construction suggestions for the learner in time, so that the constructed personalized theme map is more reasonable and more in line with the specifications.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides a method and a system for detecting a topic map conflict based on knowledge association.
The invention is realized in this way, a topic map conflict detection method based on knowledge association, comprising:
method for calculating correlation R between knowledge elements in theme graph by adopting knowledge element association relation based on path and depthk；
Calculating the correlation R between the resource groups associated with the knowledge elements in the subject graphrCalculating the similarity between the knowledge elements by combining the resource description information associated with the knowledge elements;
computing correlations R between derived knowledge elements using a path and depth based approachkRelevance R calculated by utilizing word vectors constructed by resource group multi-dimensional description information labelsrCalculation of RkAnd the degree of deviation from Rr.
Further, in calculating the correlation between the knowledge elements in the subject graph based on the association relationship between the knowledge elements of the path and the depth,
formula for thematic map in structure from information theory
Wherein, common (K1, K2) represents the commonality of the knowledge elements K1 and K2 in the subject map, diff (K1, K2) represents the difference of K1 and K2 in the subject map;
common (k1, k2) and diff (k1, k2) are calculated as follows
common(k1，k2)＝γ+Depth(RCP(k1，k2))
diff(k1，k2)＝β+P(k1,k2)
Wherein P (k1, k2) ═ Rk1+ Rk2, represents the shortest path between the knowledge elements k1 and k2, γ is the Depth adjustment parameter, β is the path adjustment parameter, Depth (RCP (k1, k2)) represents the Depth of the nearest common node of k1 and k 2; the formula common (k1, k2) ═ γ + Depth (RCP (k1, k2)), diff (k1, k2) ═ β + P (k1, k2) are substituted into the formula
Obtained by
Further, calculating the relevance R among the resource groups related to the knowledge elements in the theme maprAnd calculating the similarity between the knowledge elements by combining the resource description information associated with the knowledge elements, specifically comprising:
1) subject map KG ═ C1，C2，...，Ci]i belongs to n and represents C in an organization unit of a certain knowledge elementiThe names, the description labels and the description contents are subjected to word segmentation and part-of-speech tagging, stop words are filtered, only words with specified part-of-speech are reserved, a group of phrases for describing knowledge elements is constructed, the phrases are all composed of nouns, verbs and adjectives, and the phrases exist to form Ci＝[ti，1，ti，2，...，ti，n]In which C isiSome of the representations are a set characterized by a knowledgeelement and all tag information associated with the knowledgeelement, where ti，jIs a keyword after it has been retained;
2) constructing a candidate keyword graph G (V, E), wherein V is a node set and consists of candidate keywords generated in 1, then constructing an edge between any two points by adopting a co-occurrence relation (co-occurrence), wherein the edge exists between the two nodes only when vocabularies corresponding to the two nodes appear in a window with the length of N, N represents the size of the window, and N words are co-occurred at most;
3) iteratively propagating the weight of each node according to a TextRank algorithm until convergence; the TextRank algorithm is as follows:
4) performing reverse-narrative sequencing on the node weights to obtain m most important keywords as representation keywords;
by extracting the characterizing knowledgebase KiThe keyword set TC obtained after the keyword abstracti＝[ti，1，ti，2，...，ti，m](ii) a If calculating the element of knowledge KiAnd KjThen calculating the keyword set TCi、TCjSimilarity of (c);
similarity of single keyword representations
Further, calculating the deviation consistency of Rk and Rr, comprising:
the similarity between the knowledge element groups is calculated by two different methods, namely, the first method adopts a path and depth-based method to calculate the similarity between the knowledge element groups and the second method adopts the similarity between resource groups (including the knowledge elements) associated with the knowledge elements, and the similarity between the knowledge elements is calculated by combining the resource description information associated with the knowledge elements to obtain the knowledge elements K in the theme mapiAnd element of knowledge KjSimilarity between the two knowledge elements, if the relative position deviation of the two knowledge elements in the structure of the subject map is large, the similarity compared by the knowledge element and resource group representation is constructed according to the subject map, and the ratio psi of the two final similarities tends to a value near 1;
rk and Rr respectively represent the correlation between the knowledge elements obtained by calculating based on the way of the path and the depth and the correlation between the resource groups associated with the knowledge elements in the calculation subject map. The overall deviation factor is denoted by.
Another object of the present invention is to provide a computer program for implementing the method for detecting conflict of topic map based on knowledge association.
Another object of the present invention is to provide an information data processing terminal for implementing the knowledge association-based theme map conflict detection method.
Another object of the present invention is to provide a computer-readable storage medium, which includes instructions that, when executed on a computer, cause the computer to execute the method for detecting a topic map conflict based on knowledge correlation.
Another object of the present invention is to provide a topic map conflict detection system based on knowledge association, comprising:
a correlation calculation module among the knowledge elements in the theme map calculates the correlation R among the knowledge elements in the theme map by adopting a method of knowledge element association relation based on path and depthk；
A correlation calculation module among the resource groups, which calculates the correlation R among the resource groups related by the knowledge elements in the subject maprCalculating the similarity between the knowledge elements by combining the resource description information associated with the knowledge elements;
a deviation degree consistency calculation module for calculating the correlation R between the obtained knowledge elements by using a way based on the path and the depthkRelevance R calculated by utilizing word vectors constructed by resource group multi-dimensional description information labelsrCalculation of RkAnd the degree of deviation from Rr.
The invention also aims to provide an education platform carrying the theme map conflict detection system based on knowledge association.
Another object of the present invention is to provide a computer equipped with the theme map collision detection system based on knowledge association.
In summary, the advantages and positive effects of the invention are:
the invention mainly focuses on solving the conflict detection of the incidence relation in the construction of the personalized theme map with relatively less constructed nodes. By using the method to detect the constructed subject map, whether the relationship is reasonable or not can be quickly identified. The theme graph is an incidence relation constructed based on a tree-shaped data structure, and is more beneficial to visual presentation and maintenance of the incidence relation compared with the relation constructed by a knowledge graph graphic data structure. The following diagram is a presentation comparing the construction of a subject graph using a tree structure and the construction of an association relationship using a knowledge graph schema constraint.
The method for detecting the topic map conflict based on knowledge association can enable the actually constructed knowledge system to be more reasonable corresponding to the learning depth and the cognitive breadth of the learner, and reduce the defined deviation; the existence of all related knowledge elements under the knowledge element group keeps consistency; meanwhile, the scope of knowledge represented by most knowledge elements and a small part of knowledge elements is kept below the description scope of the same knowledge element group. The method is beneficial to enhancing the knowledge structure combing and the relevance of the resources related to knowledge, and is beneficial to better and faster construction of a personalized knowledge system by learners.
Drawings
Fig. 1 is a flowchart of a topic map conflict detection method based on knowledge association according to an embodiment of the present invention.
Fig. 2 is a schematic diagram of a theme map structure according to an embodiment of the present invention.
Fig. 3 is a schematic structural diagram of a single knowledge element and its associated resource in a topic map provided by an embodiment of the present invention.
Fig. 4 is a schematic structural diagram of an organization unit for calculating similarity in a subject map according to an embodiment of the present invention.
Fig. 5 is a schematic diagram of an association relationship between two knowledge elements k1 and k2 according to an embodiment of the present invention.
FIG. 6 is a schematic diagram of branch spacing between two knowledge elements in a nearest common parent node according to an embodiment of the present invention.
Fig. 7 is a schematic diagram of a topic map conflict detection system based on knowledge association according to an embodiment of the present invention.
In the figure: 1. a correlation calculation module among knowledge elements in the theme map; 2. a correlation calculation module among the resource groups; 3. and a deviation degree consistency calculation module.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is further described in detail with reference to the following embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
The following detailed description of the invention refers to the accompanying drawings.
As shown in fig. 1, the method for detecting a topic map conflict based on knowledge association provided by the present invention includes:
s101: calculating the correlation R between knowledge elements in the theme mapkThe method is based on the calculation of the association relation of the path and the depth knowledge elements;
s102: calculating the correlation R between resource groups (containing knowledge elements) associated with the knowledge elements in the subject graphrCalculating the similarity between the knowledge elements by combining the resource description information associated with the knowledge elements;
s103: calculating the deviation consistency of Rk and Rr, and calculating the correlation R between the obtained knowledge elements by using a path and depth-based modekRelevance R calculated by utilizing word vectors constructed by resource group multi-dimensional description information labelsrThe consistency of the two in the degree of deviation is calculated.
In step S101, the invention provides a method for calculating the relevance R of any knowledge element in the subject mapkThe method is based on the calculation of the association relation of the path and the depth knowledge elements. In the semantic knowledge element library,the knowledge elements are leaf nodes of a semantic knowledge element base, and may or may not be at the same level depth. The mode of the association relationship of the knowledge elements is undirected, starting from one knowledge element A, another knowledge element B can be found through a path, and similarly, starting from one knowledge element B, the knowledge element A can be found. This way there may not be only one path from one knowledgeelement to another, but the depth may be different. If the knowledge element belongs to the node with the public knowledge element group, the similarity calculation can be carried out by depending on the path and the depth structure mode of the subject map. Based on the structure layers, the method calculates the incidence relation and the incidence strength between two knowledge elements, and calculates the correlation of the knowledge elements close to a leaf node layer in a theme graph tree structure by adopting a mode based on the path depth and the commonality of the knowledge elements;
in step S102, the invention provides a method for calculating the correlation R between resource groups (including knowledge elements) associated with knowledge elements in a subject graphrThere is also a semantic method to calculate any two knowledge elements K in a topic map KGi，KjThe method proposed in the present invention for calculating the similarity thereof is stated as follows: before the subject map is constructed, each knowledge element has some resources related to the corresponding most relevance, knowledge element KiIs associated with R1，R2，R3...RmThe resource of (2). Calculating the similarity between the knowledge elements by combining the resource description information associated with the knowledge elements;
in step S103, the invention calculates the deviation consistency of Rk and Rr, and calculates the correlation R between the obtained knowledge elements by using a way based on the path and the depthkRelevance R calculated by utilizing word vectors constructed by resource group multi-dimensional description information labelsrThe consistency of the two on the deviation degree is calculated by representing that the knowledge elements and the resource groups share one common father node, and the knowledge content in the same range is represented in the constructed knowledge structure, so even if the knowledge elements have large differenceThere is also a deviation in consistency from the variability between the identities and the resource groups with which they are associated. Perhaps because the content and range represented by the asset description information tag have slight deviation, but the final deviation threshold value should be within a certain range on the whole, and if the set threshold value is exceeded, the learner can be prompted to have some unreasonable property for constructing the personalized theme map.
As shown in fig. 2, the subject graph structure diagram provided by the present invention includes: subject map: the theme map is a description representation of a tree structure related to a class of knowledge, which is abstracted from the upward direction of the knowledge element group, has the characteristic of individuation, and comprises at least three layers of relations, namely a resource layer, a knowledge element layer and a knowledge element group layer (or a theme layer) from the lower direction to the upper direction; knowledge element: the minimum unit which can not be subdivided and is independent in the subject map and can completely express knowledge has atomicity; semantic knowledge meta-base: and characterizing a resource library for storing the knowledge elements according to a semantic level.
Resources in the topic map have more perfect names and labels describing information: the dimension specification of the resource description comprises title names, keywords, resource content summarization (mainly comprising noun labels of clear concepts and concise summarizations), and resource attribute information (attributes such as types and sizes). The description dimension of the knowledge element or the knowledge tuple comprises five dimensions of domain, discipline, level range, allelic knowledge element or tuple and knowledge meaning to represent the knowledge details represented by the semantic knowledge element and the tuple.
As shown in fig. 3, the present invention provides a schematic structural diagram of a single knowledge element and its associated resources in a topic map.
As shown in fig. 4, the organization unit structure diagram for calculating similarity in the subject chart provided by the present invention is shown.
The embodiment of the invention also provides a method for calculating any two knowledge elements K in a theme map KG in a semantic modei，KjThe method proposed in this patent for calculating the similarity thereof is stated as follows: before the topic map is constructed, each of the knowledge elements has some resources associated with the corresponding most relevance,the relation of the knowledge element K as shown in FIG. 4iIs associated with R1，R2，R3...RmThe resource of (2). If the following two knowledge elements K are calculatedi，KjThe method for similarity of (2) is to use the knowledge element KiTo illustrate, 1. finishing KiName of (2), descriptive tag information, and resource R associated therewith1，R2，R3...R m2, extracting the most characterized knowledge element K in the label information of the names and the descriptions by using a TextRank algorithmiThe first m key words of (1) form a representation knowledge element KiDescription set TC ofi＝[ti，1，ti，2，...，ti，m]In the same way, the characteristic knowledge element K can be obtainedjDescription set TC ofj＝[tj，1，tj，2，...，tj，m](ii) a 3. And carrying out many-to-many similarity calculation by using the obtained word vectors of the description set, and then calculating the similarity of the keywords of the whole description set. In the above, the similarity between the knowledge elements is calculated by combining the resource description information associated with the knowledge elements.
The application of the present invention will be further described with reference to the following examples.
The embodiment of the invention provides the relevance R among the knowledge elements in the calculation subject chartkThe method comprises the following steps:
and calculating the correlation among the knowledge elements in the theme graph based on the association relationship of the knowledge elements of the path and the depth. The knowledge elements in the constructed subject map are the minimum units of knowledge, and different knowledge units can be formed in different arrangement and combination modes. In the semantic knowledge element library, the association mode of the knowledge elements is undirected, and the knowledge element A can be found through a path from the knowledge element A. There may not be only one path from one element to another, and the depth may be different. If the knowledge elements have a certain association relationship, the association relationship between the knowledge elements can be represented by the commonality of the path and the depth by utilizing a knowledge element association relationship calculation method based on the path and the depth. Meanwhile, based on the semantic knowledge element base, the knowledge elements are leaf nodes in the semantic knowledge element base, and may or may not be at the same level depth. According to the characteristics of the knowledge elements, the invention can make the following definitions and inferences about the relationship between the knowledge elements:
defining one: the association weights of the knowledge elements under the same knowledge element group are equal;
deducing one: associating a knowledge element with a plurality of knowledge elements under the knowledge element group, wherein the association weights are equal;
and deducing two: the association weights between the different levels of knowledge elements and the same level of knowledge elements are not equal.
In the semantic knowledgebase, for any two knowledgeable elements k1 and k2, their association relationship can be abstracted as shown in fig. 5.
As shown in fig. 5, in combination with the structural features of the topic map, the similarity between two objects is calculated in a way based on the path and the depth, and the topic map in the structure is formulated as follows according to the commonality and the difference between the two objects
Wherein common (K1, K2) in formula 1 indicates the commonality of K1 and K2 in the subject map, and diff (K1, K2) indicates the difference between K1 and K2 in the subject map.
The calculation formulas based on the structure of the above abstract path and common (k1, k2) and diff (k1, k2) described in formula 1 are the following formulas 2, 3
common (k1, k2) ═ γ + Depth (RCP (k1, k2)) formula 2
diff (k1, k2) ═ β + P (k1, k2) formula 3
Where P (k1, k2) ═ Rk1+ Rk2, represents the shortest path between the knowledge elements k1 and k2, γ is the Depth adjustment parameter, β is the path adjustment parameter, and Depth (RCP (k1, k2)) represents the Depth of the nearest common node of k1 and k 2. The result of substituting the above equations 2 and 3 into equation 1 is
In the semantic knowledge element library, common father nodes under different large categories under the same discipline are disciplines, the depth of a root node is 0, and the weight of knowledge association at the same level under different large categories is avoided to be 0. The value interval of gamma is generally (0, 1).
As shown in fig. 6, in the semantic knowledgebase, nodes follow the principle from abstraction to general from top to bottom, and two nodes with closer distances have closer meanings, so that the association degree of two knowledges is linearly and negatively correlated with the branch distance of the two knowledges in the nearest common parent node, that is, the difference between the numbers M of the directly associated knowledges or knowledges under each knowledges group is not much the same, so that the relative distance between two knowledges at the branch level is taken as the path adjustment parameter β and is taken as the extension of the path between two knowledges.
Equation 5, where
Where M is the number of direct children of the nearest common parent node of two elements or tuples, and K represents the branch distance between two elements in the nearest common parent node, as shown in the following figure, K is 2, and M is 5 between elements K7 and K8.
In the method for calculating the correlation Rr between resource groups (including the knowledge elements) associated with the knowledge elements in the subject map provided by the embodiment of the invention, the resource groups associated with the knowledge elements in the subject map include organization units formed by the knowledge elements, and then the similarity between the resource groups is calculated by a calculation method mainly depending on semantics. The premise is that both the knowledge elements and the resources have a relatively perfect descriptive information label system. An organization unit is formed by knowledge elements and resource groups associated with the knowledge elements, and the organization unit comprises two types of description information about the knowledge elements, wherein one type of description label is a description label of the knowledge elements, and the other type of description label is a resource information description label of the resource groups associated with the knowledge elements, so that a TextRank algorithm idea can be used for obtaining the description information for representing the knowledge elements K according to the following stepsiThe key word abstract of (1):
1) subject map KG ═ C1，C2，...，Ci]i belongs to n and represents C in an organization unit of a certain knowledge elementiThe names, the description labels and the description contents are subjected to word segmentation and part-of-speech tagging, stop words are filtered, only words with specified part-of-speech are reserved, a group of phrases for describing knowledge elements is constructed, the phrases are all composed of nouns, verbs and adjectives, and the phrases exist to form Ci＝[ti，1，ti，2，...ti，n]In which C isiSome of the representations are a set characterized by a knowledgeelement and all tag information associated with the knowledgeelement, where ti，jIs a keyword after it has been retained;
2) and constructing a candidate keyword graph G which is (V, E), wherein V is a node set and consists of candidate keywords generated in 1, then constructing an edge between any two points by adopting a co-occurrence relation (co-occurrence), wherein the edge exists between the two nodes only when the corresponding words appear in a window with the length of N, and N represents the size of the window, namely, at most N words co-occur.
3) Iteratively propagating the weights of the nodes until convergence according to the TextRank algorithm, equation 6, wherein w, which is used to represent the weights of the edges of all words, since keyword extraction is performed on numerous information tags using the TextRank algorithmjiAnd wjkThe values of (a) are the same, there is no intersection, there is no similarity;
4) performing reverse-narrating sequencing on the node weights to obtain m most important keywords as representation keywords;
by extracting the characterizing knowledgebase KiThe keyword set TC obtained after the keyword abstracti＝[ti，1，ti，2，...ti，m](ii) a If necessary, computing the element of knowledge KiAnd KjCalculating the similarity of the keyword sets TCi、TCjThe similarity of (c).
Similarity of single keyword representations
The deviation consistency of Rk and Rr is calculated in the embodiment of the invention;
the knowledge element K in the theme map calculated by two different methodsiAnd element of knowledge KjIf the relative position deviation of the two knowledge elements in the structure of the subject map is large, the deviation of the similarity compared by the knowledge element and resource group representation according to the way of building the subject map is also large, and the ratio psi of the two obtained final similarities tends to be a value near 1.
As shown in fig. 7, the topic map conflict detection system based on knowledge association provided in the embodiment of the present invention includes:
the correlation calculation module 1 between the knowledge elements in the theme map calculates the correlation R between the knowledge elements in the theme map by adopting a method of the association relationship of the knowledge elements based on the path and the depthk；
A correlation calculation module 2 among the resource groups, which calculates the correlation R among the resource groups associated by the knowledge elements in the subject graphrCalculating the similarity between the knowledge elements by combining the resource description information associated with the knowledge elements;
a deviation degree consistency calculation module 3 for calculating the correlation R between the obtained knowledge elements by using a way based on the path and the depthkRelevance R calculated by utilizing word vectors constructed by resource group multi-dimensional description information labelsrCalculation of RkAnd the degree of deviation from Rr.
In the above embodiments, the implementation may be wholly or partially realized by software, hardware, firmware, or any combination thereof. When used in whole or in part, can be implemented in a computer program product that includes one or more computer instructions. When loaded or executed on a computer, cause the flow or functions according to embodiments of the invention to occur, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a network of computers, or other programmable device. The computer instructions may be stored in a computer readable storage medium or transmitted from one computer readable storage medium to another, for example, the computer instructions may be transmitted from one website site, computer, server, or data center to another website site, computer, server, or data center via wire (e.g., coaxial cable, fiber optic, Digital Subscriber Line (DSL), or wireless (e.g., infrared, wireless, microwave, etc.)). The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device, such as a server, a data center, etc., that includes one or more of the available media. The usable medium may be a magnetic medium (e.g., floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., Solid State Disk (SSD)), among others.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (6)
1. A topic map conflict detection method based on knowledge association is characterized by comprising the following steps:
method for calculating correlation between knowledge elements in theme graph by adopting knowledge element association relation based on path and depth
；
Calculating the correlation among the resource groups associated with the knowledge elements in the subject graph
Calculating the similarity between the knowledge elements by combining the resource description information associated with the knowledge elements;
computing correlations between derived knowledge elements using a path and depth based approach
Correlation calculated by using word vector constructed by resource group multi-dimensional description information label
CalculatingR kConsistency with Rr in degree of deviation;
in calculating the correlation between the knowledge elements in the theme map based on the association relationship between the knowledge elements of the path and the depth,
formula for thematic map in structure from information theory
Wherein,
the commonality of the represented knowledge elements K1 and K2 in the subject graph, diff (K1, K2) represents the difference between K1 and K2 in the subject graph;
Wherein P (k1, k2) = Rk1+ Rk2, which represents the shortest path between the knowledge elements k1 and k2, γ is a depth adjustment parameter, β is a path adjustment parameter,
represents the depth of the nearest common node of k1 and k 2; will be a formula
=γ+
、
Carry over of = β + P (k1, k2) to formula
Obtained by
Calculating the correlation among the resource groups associated with the knowledge elements in the subject graph
And calculating the similarity between the knowledge elements by combining the resource description information associated with the knowledge elements, specifically comprising:
1) subject map KG =
In an organization unit characterizing a knowledge element
The names, the description labels and the description contents are subjected to word segmentation and part-of-speech tagging, stop words are filtered, only words with specified part-of-speech are reserved, a group of phrases for describing knowledge elements is constructed, the phrases are all composed of nouns, verbs and adjectives, and the phrases exist to form the phrases
Wherein
Some of the representations are a set characterized by a knowledgeelement and all tag information associated with the knowledgeelement, where
Is a keyword after it has been retained;
2) constructing a candidate keyword graph G = (V, E), wherein V is a node set and consists of candidate keywords generated in 1, then constructing an edge between any two points by adopting a co-occurrence relation (co-occurrence), the edge between the two nodes only appears when vocabularies corresponding to the two nodes appear in a window with the length of N, N represents the size of the window, and N words coexist at most;
3) iteratively propagating the weight of each node according to a TextRank algorithm until convergence; the TextRank algorithm is as follows:
4) performing reverse-narrative sequencing on the node weights to obtain m most important keywords as representation keywords;
characterizing knowledge elements by extraction
The key word set obtained after the key word abstract
(ii) a If calculating the knowledge element
And
then calculate the keyword set
、
Similarity of (c);
similarity of single keyword representations
2. The knowledge correlation based topic map collision detection method of claim 1,
calculating the deviation consistency of Rk and Rr, comprising:
combining the similarity between the knowledge elements by adopting two methods of calculating the similarity between the knowledge element groups based on the path and the depth and adopting the similarity between the resource groups associated with the knowledge elementsResource description information, and the subject map obtained by calculating the similarity between the knowledge elements
And the element of knowledge
Similarity between the knowledge elements is large in relative position deviation of the two knowledge elements in the structure of the subject map, and the final two similarity values are obtained through similarity comparison of the knowledge elements and resource group representations according to the subject map construction mode
Values approaching around 1;
Rkand RrRespectively representing the correlation between the knowledge elements obtained by calculation based on the way of the path and the depth and the correlation between the resource groups associated with the knowledge elements in the calculation subject map; the overall deviation factor is denoted by.
3. An information data processing terminal for implementing the knowledge association-based topic map conflict detection method of any one of claims 1 to 2.
4. A computer-readable storage medium comprising instructions that, when executed on a computer, cause the computer to perform the method for topic map conflict detection based on knowledge correlation of any one of claims 1-2.
5. A knowledge association based topic map conflict detection system for implementing the knowledge association based topic map conflict detection method of claim 1, wherein the knowledge association based topic map conflict detection system comprises:
a correlation calculation module among the knowledge elements in the theme map calculates the correlation among the knowledge elements in the theme map by adopting a method of knowledge element association relation based on path and depth
；
A correlation calculation module among the resource groups for calculating the correlation among the resource groups associated with the knowledge elements in the subject map
Calculating the similarity between the knowledge elements by combining the resource description information associated with the knowledge elements;
a deviation degree consistency calculation module for calculating the correlation between the obtained knowledge elements by using a way based on the path and the depth
Correlation calculated by using word vector constructed by resource group multi-dimensional description information label
ComputingR kAnd the degree of deviation from Rr.
6. A computer loaded with the knowledge association based topic map collision detection system of claim 5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810918417.4A CN109344293B (en) | 2018-08-13 | 2018-08-13 | Knowledge association-based theme map conflict detection method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810918417.4A CN109344293B (en) | 2018-08-13 | 2018-08-13 | Knowledge association-based theme map conflict detection method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109344293A CN109344293A (en) | 2019-02-15 |
| CN109344293B true CN109344293B (en) | 2021-05-18 |
Family
ID=65291707
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810918417.4A Active CN109344293B (en) | 2018-08-13 | 2018-08-13 | Knowledge association-based theme map conflict detection method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109344293B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114547259B (en) * | 2020-11-26 | 2024-05-24 | 北京大学 | Automatic formula description generation method and system based on topic relation graph |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20150033768A (en) * | 2013-09-23 | 2015-04-02 | 충북대학교 산학협력단 | System and method for expert search by dynamic profile and social network reliability |
| CN105335374A (en) * | 2014-06-19 | 2016-02-17 | 北大方正集团有限公司 | Knowledge point association method and apparatus as well as server and client containing apparatus |
| CN105893362A (en) * | 2014-09-26 | 2016-08-24 | 北大方正集团有限公司 | A method for acquiring knowledge point semantic vectors and a method and a system for determining correlative knowledge points |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9110882B2 (en) * | 2010-05-14 | 2015-08-18 | Amazon Technologies, Inc. | Extracting structured knowledge from unstructured text |
| US20180052884A1 (en) * | 2016-08-16 | 2018-02-22 | Ebay Inc. | Knowledge graph construction for intelligent online personal assistant |
-
2018
- 2018-08-13 CN CN201810918417.4A patent/CN109344293B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20150033768A (en) * | 2013-09-23 | 2015-04-02 | 충북대학교 산학협력단 | System and method for expert search by dynamic profile and social network reliability |
| CN105335374A (en) * | 2014-06-19 | 2016-02-17 | 北大方正集团有限公司 | Knowledge point association method and apparatus as well as server and client containing apparatus |
| CN105893362A (en) * | 2014-09-26 | 2016-08-24 | 北大方正集团有限公司 | A method for acquiring knowledge point semantic vectors and a method and a system for determining correlative knowledge points |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109344293A (en) | 2019-02-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| KR20200094627A (en) | Method, apparatus, device and medium for determining text relevance | |
| WO2018205838A1 (en) | Method and apparatus for retrieving similar video, and storage medium | |
| CN109804371B (en) | Method and device for semantic knowledge migration | |
| CN111522886B (en) | Information recommendation method, terminal and storage medium | |
| US10614138B2 (en) | Taste extraction curation and tagging | |
| WO2014210387A2 (en) | Concept extraction | |
| CN108874783A (en) | Power information O&M knowledge model construction method | |
| CN109992784B (en) | Heterogeneous network construction and distance measurement method fusing multi-mode information | |
| CN110275962B (en) | Method and apparatus for outputting information | |
| CN111797222A (en) | Course knowledge graph construction method, device, terminal and storage medium | |
| CN106897437B (en) | High-order rule multi-classification method and system of knowledge system | |
| CN110851609A (en) | Representation learning method and device | |
| US20100138414A1 (en) | Methods and systems for associative search | |
| CN109002432A (en) | Method for digging and device, computer-readable medium, the electronic equipment of synonym | |
| CN106777218B (en) | Ontology matching method based on attribute similarity | |
| CN111475653B (en) | Method and device for constructing knowledge graph in oil and gas exploration and development field | |
| CN114840685A (en) | Emergency plan knowledge graph construction method | |
| CN109344293B (en) | Knowledge association-based theme map conflict detection method and system | |
| CN113268607A (en) | Knowledge graph construction method and device | |
| Xu et al. | A novel entity joint annotation relation extraction model | |
| CN103136221A (en) | Method capable of generating requirement template and requirement identification method and device | |
| CN116610874A (en) | Cross-domain recommendation method based on knowledge graph and graph neural network | |
| CN114201622B (en) | Method and device for acquiring event information, electronic equipment and storage medium | |
| CN104750692B (en) | A kind of information processing method, information retrieval method and its corresponding device | |
| Maake et al. | A bisociated research paper recommendation model using BiSOLinkers |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |