KR100947579B1 - Method and apparatus for measuring semantic distance between concepts in concept tree - Google Patents

Abstract

PURPOSE: A method and an apparatus for measuring a semantic distance among concepts in a concept tree are provided to calculate a common ancestor node between two concepts using a virtual identification number assigned to each concept in a concept tree. CONSTITUTION: An identification number assigning unit(902) assigns an identification number to a node included in a concept tree. A virtual identification number assigning unit(904) converts the concept tree into a complete tree shape using the maximum child number information of the concept tree. The virtual identification number assigning unit assigns a virtual identification number to the node. A distance calculating unit(906) calculates a distance between two nodes set among nodes included in the concept tree using the virtual identification number.

Description

Method and apparatus for measuring semantic distance between concepts in concept tree}

Embodiments of the invention relate to techniques for measuring the difference in meaning between terms in an ontology represented in tree or graph form.

Ontology is a model that abstracts and shares what people think about things. It is a technique that is formalized and explicitly defines the type of concept or constraints on its use. In particular, as a data model representing a specific domain in computer science and information science, it is defined as a set of formal vocabulary describing concepts belonging to a specific domain and the relationships between the concepts.

In the ontology, the isA relationship between concepts is generally expressed in the form of a concept tree or graph. The semantic distance between the concepts in the ontology is the number of edges connecting the concepts that make up the ontology. Will be calculated. However, in order to calculate the number of edges on the path connecting the concepts as described above, the number of edges must be calculated while directly moving the path in the tree or graph. Or we need a way to calculate the distance between two concepts in a graph.

Embodiments of the present invention assign a virtual identification number to each concept in a concept tree and calculate a common ancestor node between any two concepts using the virtual identification number, thereby establishing a path connecting the concepts within the concept tree. It is intended to provide a method for calculating the distance between concepts by simple calculation without moving directly.

Method for measuring the semantic distance between concepts in the concept tree according to an embodiment of the present invention for solving the above problems, the semantic distance measuring apparatus, the step of assigning an identification number to each node constituting the concept tree; In the semantic distance measuring apparatus, converting the concept tree into a complete tree using information on the maximum number of children of the concept tree and assigning a virtual identification number to each of the nodes; And calculating, by the apparatus for semantic distance measurement, a distance between two preset nodes among nodes included in the concept tree using the virtual identification number.

Meanwhile, an apparatus for measuring semantic distance between concepts in a concept tree according to an embodiment of the present invention for solving the above problems includes an identification number allocator for allocating an identification number to each node included in the concept tree; A virtual identification number allocator for converting the concept tree into a complete tree using information on the maximum number of children of the concept tree and allocating a virtual identification number to each of the nodes; And a distance calculator configured to calculate a distance between two preset nodes among the nodes included in the concept tree using the virtual identification number.

Other aspects, features, and advantages other than those described above will become apparent from the following drawings, claims, and detailed description of the invention.

Embodiments of the present invention calculate a common ancestor node between any two concepts using a virtual identification number assigned to each concept in the concept tree, without directly moving the path connecting the concepts within the concept tree. By simple calculation, the distance between each concept can be calculated.

Hereinafter, specific embodiments of the present invention will be described with reference to the drawings. However, this is only an example and the present invention is not limited thereto.

In describing the present invention, when it is determined that the detailed description of the known technology related to the present invention may unnecessarily obscure the subject matter of the present invention, the detailed description thereof will be omitted. In addition, terms to be described below are terms defined in consideration of functions in the present invention, which may vary according to the intention or custom of a user or an operator. Therefore, the definition should be made based on the contents throughout the specification.

The technical spirit of the present invention is determined by the claims, and the following embodiments are merely one means for effectively explaining the technical spirit of the present invention to those skilled in the art.

Prior to describing the present invention, a concept tree according to an embodiment of the present invention will be briefly described.

1 is a diagram illustrating an example of a concept tree 100 according to an embodiment of the present invention. In the embodiment of the present invention, the isA relationship of the ontology is configured in a tree form, and the semantic distance between the nodes is expressed as the number of edges between the nodes of the tree. For example, in the concept tree 100 illustrated in FIG. 1, the distance between the "animal" and the "invertebrate" is 1, the distance between the "invertebrate" and the "vertebrate" is 2, and the "annular animal" and the "fish". The distance is four. In the following description, "distance" refers to the semantically distance (or semantic distance).

In the embodiment of the present invention, the concept tree is configured in the form of a K-ary tree. K-ary tree is a kind of rooted tree, which means a tree with a maximum number of children of each node of the tree. For example, in the concept tree 100 shown in FIG. 1, since the number of children of the "vertebrate" is the most, 3, K becomes 3.

In the rooted tree, as shown in Fig. 1, when the diagram is configured to have a root at the top and a path downward, a node immediately below any node is a child node, and a node immediately above is a parent node. (parent node) For example, in FIG. 1 the parent node of "invertebrate" is "animal", child nodes are "annular animals" and "mollusks". In addition, all nodes in the upstream path from a node to the root are ancestors of all nodes, all nodes in the down path are descendants, and multiple nodes with the same parent are synchronized with each other. sibling) relationship. There is a sequence between each sync node, for example the sequence between "float", "bird", "fish" can be defined as 1, 2, and 3, respectively.

At the top of the concept tree, the root, is the root node. The root node is a node with no parent and only children. On the other hand, a node with no children and only a parent in the concept tree is called a leaf node. All nodes except the root and leaf nodes have parent and child nodes.

For any node U in a rooted tree, the length of the path from the root to the node U, ie the number of nodes on the path, is called the level or depth of U. . For example, the level of the root node is 1, and the level of child nodes of the root node is 2. The level of the root node may start with 0 or another number other than 1, but the level number is assigned so that the level difference between the parent node and the child node is 1.

2 is a flowchart illustrating a method 200 for measuring semantic distance between concepts in a concept tree according to an embodiment of the present invention.

As shown, a method 200 for measuring semantic distance between concepts in a concept tree in accordance with an embodiment of the present invention includes assigning an identification number to the concept tree (202), and adding a virtual node to the concept tree. Converting to a full K-ary tree (204), assigning a virtual identification number to the concept tree using the concept tree converted to the full K-ary tree (206), and the identification number and virtual identification Generating an identification number translation table including the number (208) and calculating a distance between nodes using the identification number translation table (210).

Hereinafter, the semantic distance measuring method 200 between the concepts in the concept tree according to an embodiment of the present invention will be described in detail for each step. For convenience of explanation, the concept tree shown in FIG. 1 will be described below as an example.

Assigning Identification Numbers to the Concept Tree (202)

In this step, an identification number for identifying each concept (node) constituting the concept tree is assigned. The identification number may be sequentially assigned, for example, starting with the identification number of the root node of the concept tree starting with 1 and according to the level of the concept tree. In addition, in the case of nodes of the same level, identification numbers may be sequentially assigned according to the sequence of the nodes. An embodiment in which an identification number is assigned in this manner is shown in FIG. 3. However, in the present invention, it should be noted that there is no particular limitation on the method of allocating the identification number, and any method of assigning a unique identification number to each sensor node may be used.

Concept tree full K- ary  Convert to Tree (204)

In this step, the concept tree is converted into a complete K-ary tree in order to assign a virtual identification number separate from the actual identification number to each node of the concept tree to which the identification number is assigned.

First, a maximum number of children (K) is calculated in the concept tree, and a virtual node is added to the concept tree to convert it into a perfect K-ary tree. The virtual node is not an actual node, but a virtual node used for assigning a virtual identification number. A perfect tree is a tree in which all nodes except the highest level nodes have as many children as the maximum number of children (K). For example, in the concept tree illustrated in FIG. 3, K = 3, but in the case of nodes 1 and 2, the number of child nodes is two, so it is not a perfect tree. Add a virtual node so that all nodes have the maximum number of children, and transform the leaf nodes so that they exist only at the highest level, resulting in a complete K-ary tree. On the other hand, a complete tree is a complete tree where leaf nodes exist only at the largest level h, or leaf nodes may exist at level h and level h-1, and up to level h-1 in the form of a complete tree, Leaf nodes at level h are filled sequentially from the left to the last node. This type of K-ary tree may be referred to as a complete K-ary tree. In the present invention, this is called a complete tree for convenience.

Adding virtual nodes such that all nodes at level 1 and level 2 of FIG. 3 have maximum child nodes (the virtual nodes in the figure are indicated by dotted lines) results in a complete tree as shown in FIG. Also, in FIG. 4, adding virtual nodes such that all nodes (including virtual nodes) of level 2 have a maximum number of children becomes a perfect K-ary tree.

Assigning Virtual Identification Numbers to the Concept Tree (206)

Next, the virtual node is assigned a virtual identification number in the same level order as when the virtual node assigns the identification number to the added complete tree. The virtual identification number assigned in this way is shown in each node of FIG. In the figure, the left side of ":" represents an identification number and the right side represents a virtual identification number. For example, a node having "8:10" means a node having an identification number of 8 and a virtual identification number of 10. In the case of the virtual node, since no identification number exists, only the virtual identification number is described.

Create Identification Number Translation Table (208)

When the identification number and the virtual identification number are assigned to all nodes of the concept tree, the identification number translation table including the information of each node (identification number, virtual identification number) is stored and managed. The identification number translation table is a table in which identification number and virtual identification number pairs of all nodes of the concept tree are stored.

Table 1 below shows an example of the identification number conversion table according to the embodiment of the present invention described above.

Physical identification number Virtual identification number One One 2 2 3 3 4 5 5 6 6 8 7 9 8 10

When the identification number translation table is generated as described above, the parent node and the child node of the node can be known from the specific node of the concept tree.

First, when the maximum number of children of the concept tree is K and the virtual identification number of any node is m, the virtual identification number of the parent node of the node m is determined as in Equation 1 below.

At this time, parent (m) is a virtual identification number of the parent node of node m.

After calculating the virtual identification number of the parent node using Equation 1, the identification number of the parent node corresponding to the virtual identification number can be known using the identification number conversion table.

On the other hand, when the maximum number of children of the concept tree is K and the virtual identification number of any node is m, the virtual identification number of the j-th child node of the node m is determined as in Equation 2 below.

At this time, child (m, j) is a virtual identification number of the j-th child node of the node m.

After calculating the virtual identification number of the child node using Equation 2, the identification number of the child node corresponding to the virtual identification number can be known using the identification number conversion table.

In the concept tree, the parent node of each node represents the parent concept of the node and the child nodes represent the child concepts. In other words, there is an "isA" relationship between child-parent nodes of the concept tree. Therefore, by calculating the parent node and the child node of the node in the above manner, it is possible to know the upper and lower conceptual relationships in the concept tree of the specific concept.

Meanwhile, when the virtual identification number of a specific node is known in the concept tree, the level of the node can be calculated.

When the maximum number of children of the concept tree is K, the maximum number of nodes up to level L of the concept tree is expressed by Equation 3 below.

Max_nodes (L) is the maximum number of nodes up to level L of the concept tree.

According to Equation 3, if the level of the node having the virtual identification number m is L, it can be seen that m and L have the following relationship.

Therefore, for the node having the virtual identification number m, the L value satisfying the above expression (4) becomes the level of the node.

Calculate distance between nodes (210)

The semantically distance between any two nodes in the concept tree is defined as the number of edges connecting the two nodes, as described above. Hereinafter, a method of calculating the semantic distance between two arbitrary nodes in the concept tree will be described.

First, the virtual identification number of the i-th ancestor node of the node whose virtual identification number is m will be defined as P ⁱ (m).

Since the 0th parent node of the node having the virtual identification number m is itself, P ⁰ (m) = 1. P ¹ (m) represents its parent node, and is determined as follows according to Equation 1 described above.

P ² (m) represents the parent node of its parent node, which is expressed as follows.

Generalizing this, the virtual identification number of the i-th ancestor node of the node having the virtual identification number m is represented by Equation 5 below.

If the virtual identification number of the ancestor node of the node with the virtual identification number m and the virtual identification number of the ancestor node of the node with the virtual identification number t in the concept tree match, the common ancestor node is satisfied.

In other words, if the previous expression is satisfied, the i-th ancestor node of the node having the virtual identification number m and the j-th ancestor node of the node having the virtual identification number t become common ancestor nodes (where m and t are not the same). . One or more common ancestor nodes exist from the nearest common ancestor node (an ancestor node having the highest level value among the common ancestors of the nodes) to the root node.

 Therefore, the distance between two nodes can be known by obtaining the closest common ancestor node of the node having the virtual identification number of m and the node of t and adding the distances from each node to the common ancestor node.

That is, as shown in Equation 6 below, when the node having the virtual identification number m and the node having t satisfy the following relationship,

The distance between two nodes is (i + j). In the above equation, min (i) and min (j) mean the values of pairs i and j that are the minimum among pairs of values of i and j that satisfy the equation on the right, respectively. This is to find the nearest common ancestor node among the common ancestor nodes.

Using the above equation, in calculating the semantic distance between two nodes in the concept tree, the distance between nodes can be calculated by simple calculation without the need to count and count the number of edges directly in the concept tree.

Calculate the path between two nodes

Using the aforementioned relationship, not only the distance of any two nodes in the concept tree can be known, but also the path between the two nodes can be easily calculated without looking at the concept tree.

For example, between node m and node t in the concept tree (m, t is a virtual identification number),

If a relationship is found, then the path between m and t

P ⁰ (m) = m, P ¹ (m), P ² (m), ..., P ^{i -1} (m), P ⁱ (m) = P ^j (t) (first joint ancestor), P ^j-1 (t), P ^{j- 2} (t), ..., P ¹ (t), P ⁰ (t) = t.

Meanwhile, between two nodes with virtual identification numbers m and a

If the relationship is established, a is the i-th direct ancestor node of m and corresponds to the higher concept of m.

Also the concepts on the path between m and a

P ⁰ (m) = m, P ¹ (m), P ² (m), ..., P ^{i -1} (m), P ⁱ (m) = a

Since the above-described relationship can be used, words corresponding to higher or lower concepts of a concept can be easily extracted.

For conceptual graphs

Ontology is not necessarily composed of a concept tree, but may also have a graph form (concept graph). The concept graph differs from the concept tree in that there can be more than one parent node for a particular node. An example of such a conceptual graph is shown in FIG. For convenience of explanation, each concept (node) of the illustrated concept graph is assigned an alphabet. The maximum number K of nodes in the conceptual graph is two.

In order to calculate the distance between nodes in the concept graph as described above, the concept graph must first be converted into a concept tree. This is done by copying a node having a plurality of parent nodes and sub-nodes of the node into subtrees of each parent node. In the illustrated embodiment, if a subtree having node E as a root node is copied and copied to a child node of node C, the concept graph may be converted into a concept tree. An embodiment converted to a tree structure as described above is shown in FIG. Although FIG. 6 illustrates the case where two parent nodes of node E are illustrated, the number of parent nodes of each node in the concept graph is not limited. Therefore, in this case, the subtree must be copied to each parent node as in the previous method. In addition, when there are a plurality of nodes having a plurality of parent nodes in the concept graph, the subtree is copied to all of the plurality of nodes as in the previous method, and all nodes in the concept graph are converted into a concept tree having only one parent node. .

When copying a subtree, you can actually copy all the nodes in the subtree. However, if the data size of a node is large, you can reduce the storage space by not copying the node but by storing a pointer to the original node in the copy node. You can also help with copying.

Next, an identification number and a virtual identification number are assigned to the converted concept graph and an identification number conversion table is generated. Since it has been described in detail above, the description thereof will be omitted. FIG. 7 illustrates a conceptual graph to which identification numbers and virtual identification numbers are assigned, and the identification number conversion table generated therefrom is shown in Table 2 below.

Physical identification number Virtual identification number One One 2 2 3 3 4 4 5 5 6 6 7 7 8 8 9 10 10 12 11 14 12 15

In the case of a concept tree generated by converting from a concept graph, a method for distinguishing whether a node is copied or not is needed. Therefore, in the present embodiment, a duplicate table is generated as shown in Table 3 separately from the identification number conversion table.

Identification number Duplicate identification number 5 6 9 10

Referring to the duplicate identification number, in the concept tree shown in FIG. You can see that it is a node.

After converting the concept graph to the concept tree as described above, the distance between nodes is calculated using Equation 6 in the same manner as the distance calculation process between the nodes in the concept tree described above. However, even though the nodes appear to be different nodes on the concept tree, they may actually be the same nodes on the concept graph, and thus the duplicate table should be referred to for distance calculation.

For example, in the concept tree of FIG. 7, since the identification number 9 node and the identification number 10 node are virtual identification numbers 10 and 12, respectively,

The relationship between the two nodes is calculated as 3 + 3 = 6. However, referring to the duplicate table, since the node with the identification number 10 is a node copied from the node with the identification number 9, the actual distance between the two nodes becomes zero.

trunk( edge ) Different distances

In the above-described embodiment, although the distance between the trunk lines is uniformly assumed to be 1 in the concept tree, the distance between the trunk lines may be determined differently according to embodiments. For example, in the concept tree of FIG. 8, the distance between nodes B and D is defined as 0.5, and the distance between A and C is defined as 2.0.

As such, when the distances are different for each edge, the distance between nodes may be calculated as follows.

For example, when calculating distances between nodes having virtual identification numbers m and t, respectively, in a conceptual tree having different distances for each trunk, first, i and j values satisfying the relationship shown in Equation 6 are calculated.

Then, the path between m and t is obtained by the method described above.

P ⁰ (m) = m, P ¹ (m), P ² (m), ..., P ^{i -1} (m), P ⁱ (m) = P ^j (t) (first joint ancestor), P ^j-1 (t), P ^{j -2} (t), ..., P ¹ (t), P ⁰ (t) = t

Finally, the distance between each node constituting the path from the concept tree and the sum of all the distances can be obtained. This is expressed as the following equation.

In the above equation, the i th ancestor of m and the j th ancestor of t are the first common ancestors, and d (a, b) represents a distance between a node having a virtual identification number a and a node having b, respectively. (Where node a and node b are child-parent relationships)

Meanwhile, not only the concept tree but also a concept graph having different distances between the edges, the distance between nodes can be known by converting the concept graph into the concept tree by the above-described method. However, in the case of the concept graph, there are a plurality of paths between nodes, so it is obvious that the distance between each node of the concept graph should be defined so that the distance between nodes is the same no matter which path is taken.

Meanwhile, an embodiment of the present invention may include a computer readable recording medium including a program for performing the methods described herein on a computer. The computer-readable recording medium may include program instructions, local data files, local data structures, etc. alone or in combination. The media may be those specially designed and constructed for the purposes of the present invention, or they may be of the kind well-known and available to those skilled in the computer software arts. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tape, optical recording media such as CD-ROMs, DVDs, magnetic-optical media such as floppy disks, and ROM, RAM, flash memory, and the like. Hardware devices specifically configured to store and execute program instructions are included. Examples of program instructions may include high-level language code that can be executed by a computer using an interpreter as well as machine code such as produced by a compiler.

9 is a diagram illustrating a semantic distance measuring apparatus 900 according to an embodiment of the present invention.

The semantic distance measuring apparatus 900 according to an embodiment of the present invention is an apparatus for performing the above-described semantic distance measuring method 200. As shown, an identification number assignment unit 902 and a virtual identification number assignment unit ( 904, a distance calculator 906, and a database 908.

The identification number assignment unit 902 assigns an identification number for distinguishing each concept (node) constituting the concept tree. As described above, the identification number may be sequentially assigned from the root node of the concept tree according to the level of the concept tree. In addition, in the case of nodes of the same level, identification numbers may be sequentially assigned according to the sequence of the nodes.

The virtual identification number allocator 904 converts the concept tree into a complete tree and assigns a virtual identification number separate from the actual identification number. It also generates an identification number translation table including the assigned identification number and virtual identification number pair. Details regarding the identification number, the virtual identification number, and the identification number conversion table have been described in detail above.

The distance calculator 906 calculates the distance between any two nodes in the concept tree to which the identification number and the virtual identification number are assigned in the identification number assignment unit 902 and the virtual identification number assignment unit 904. In addition, according to an embodiment, it may be configured to calculate a path connecting two nodes as well as a distance between two nodes.

The database 908 is a storage space in which the identification number conversion table generated by the virtual identification number assignment unit 904 is stored.

On the other hand, the semantic distance measuring apparatus 900 according to an embodiment of the present invention may further include a tree structure conversion unit (not shown) for converting the concept graph into a concept tree form as needed. In this case, the tree structure converter may be configured to generate a duplicate table from the converted concept tree and the concept graph and store it in the database 908.

Although the present invention has been described in detail with reference to exemplary embodiments above, those skilled in the art to which the present invention pertains can make various modifications to the above-described embodiments without departing from the scope of the present invention. I will understand.

Therefore, the scope of the present invention should not be limited to the described embodiments, but should be defined by the claims below and equivalents thereof.

1 is a diagram illustrating an example of a concept tree 100 according to an embodiment of the present invention.

2 is a flowchart illustrating a method 200 for measuring semantic distance between concepts in a concept tree according to an embodiment of the present invention.

3 is a diagram illustrating an example of a concept tree to which an identification number is assigned according to an embodiment of the present invention.

4 is a diagram illustrating an example of a concept tree converted into a complete tree and assigned a virtual identification number according to an embodiment of the present invention.

5 is a diagram illustrating an example of a conceptual graph according to an embodiment of the present invention.

6 illustrates an example of converting a concept graph into a concept tree according to an embodiment of the present invention.

7 is a diagram illustrating an example in which an identification number and a virtual identification number are assigned to a converted concept graph according to an embodiment of the present invention.

8 is a diagram illustrating a concept tree in which distances between concepts are set differently according to one embodiment of the present invention.

9 is a diagram illustrating a semantic distance measuring apparatus 900 according to an embodiment of the present invention.

Claims (21)

In the semantic distance measuring apparatus, assigning an identification number to each node constituting the concept tree; In the semantic distance measuring apparatus, converting the concept tree into a complete tree using information on the maximum number of children of the concept tree and assigning a virtual identification number to each of the nodes; And Calculating, by the apparatus for semantic distance measurement, a distance between two preset nodes among nodes included in the concept tree using the virtual identification number; Method for measuring semantic distance between concepts in a concept tree comprising a. The method of claim 1, Before performing the identification number assignment step, Converting the concept graph into a concept tree by copying a subtree including a node having a plurality of parent nodes among the nodes included in the concept graph and a subtree including sub-nodes of the node to child nodes of each of the plurality of parent nodes; Further comprising, the semantic distance measurement method between the concepts in the concept tree. The method of claim 1, And the identification number and the virtual identification number are sequentially assigned from the root node of the concept tree according to the level of the concept tree. The method of claim 1, And the distance between the two nodes is the number of edges connecting the two nodes in the concept tree. The method of claim 4, wherein The distance between the two nodes is calculated by calculating the common ancestor having the highest level among the common ancestors of the two nodes using the virtual identification number of the two nodes, and calculating the number of edges from each of the two nodes to the common ancestor. Being, how to measure the semantic distance between concepts in the concept tree. The method of claim 5, The two nodes are

Where m and j are the virtual identification numbers of the two nodes, and P ⁱ (m) is the virtual identification number of the i th ancestor node of the node whose virtual identification number is m, and min (i) and min (j) are respectively. I and j values of the pairs of i and j values satisfying the expression on the right) If satisfying, the distance between the two nodes is determined by i + j, semantic distance measurement method between concepts in the concept tree. The method of claim 6, The P ⁱ (m) value is represented by the following equation,

Where K is the maximum number of children in the concept tree A method of measuring the semantic distance between concepts in a concept tree. The method of claim 1, And a distance between the two nodes is a sum of distance values preset to edges connecting the two nodes in the concept tree. The method of claim 8, The distance between the two nodes is calculated using the virtual identification number of the two nodes, the common ancestor with the highest level among the common ancestors of the two nodes, and the node existing on the path from each of the two nodes to the common ancestor. And a semantic distance measurement method between concepts in the concept tree, which is calculated by summing all preset distance values between the extracted nodes. The method of claim 9, The two nodes are

Where m and j are the virtual identification numbers of the two nodes, and P ⁱ (m) is the virtual identification number of the i th ancestor node of the node whose virtual identification number is m, and min (i) and min (j) are respectively. I and j values of the pairs of i and j values satisfying the expression on the right) If, satisfies the distance between the two nodes, the following equation

(Where d (a, b) is the distance between nodes with virtual identification numbers a and b, respectively, where node a and node b are child-parent relationships) A method of measuring the semantic distance between concepts in a concept tree. A computer-readable recording medium having recorded thereon a program for executing the method according to any one of claims 1 to 10 on a computer. An identification number assignment unit for assigning identification numbers to each node included in the concept tree; A virtual identification number allocator for converting the concept tree into a complete tree using information on the maximum number of children of the concept tree and allocating a virtual identification number to each of the nodes; And A distance calculator configured to calculate a distance between two preset nodes among the nodes included in the concept tree using the virtual identification number; Semantic distance measuring device comprising a. The method of claim 12, Tree structure transformation converting the concept graph into a concept tree by copying a subtree including a node having a plurality of parent nodes and sub-nodes of the node among the nodes included in the concept graph to child nodes of each of the plurality of parent nodes. The semantic distance measuring apparatus further comprises. The method of claim 12, And the identification number and the virtual identification number are sequentially assigned from the root node of the concept tree according to the level of the concept tree. The method of claim 12, And the distance between the two nodes is the number of edges connecting the two nodes in the concept tree. The method of claim 15, The distance between the two nodes is calculated by calculating the common ancestor having the highest level among the common ancestors of the two nodes using the virtual identification number of the two nodes, and calculating the number of edges from each of the two nodes to the common ancestor. , Semantic distance measuring device. The method of claim 16, The two nodes are

Where m and j are the virtual identification numbers of the two nodes, and P ⁱ (m) is the virtual identification number of the i th ancestor node of the node whose virtual identification number is m, and min (i) and min (j) are respectively. I and j values of the pairs of i and j values satisfying the expression on the right) If satisfying, the distance between the two nodes is determined by i + j, the semantic distance measuring apparatus. The method of claim 17, The P ⁱ (m) value is represented by the following equation,

Where K is the maximum number of children in the concept tree Determined by the semantic distance measuring device. The method of claim 12, And a distance between the two nodes is a sum of distance values preset to edges connecting the two nodes in the concept tree. The method of claim 19, The distance between the two nodes is calculated using the virtual identification number of the two nodes, the common ancestor with the highest level among the common ancestors of the two nodes, and the node existing on the path from each of the two nodes to the common ancestor. Extracting and calculating the sum of the predetermined distance values between the extracted nodes. The method of claim 20, The two nodes are

Where m and j are the virtual identification numbers of the two nodes, and P ⁱ (m) is the virtual identification number of the i th ancestor node of the node whose virtual identification number is m, and min (i) and min (j) are respectively. I and j values of the pairs of i and j values satisfying the expression on the right) If, satisfies the distance between the two nodes, the following equation

(Where d (a, b) is the distance between nodes with virtual identification numbers a and b, respectively, where node a and node b are child-parent relationships) Determined by the semantic distance measuring device.

Images