EP2225676A2

EP2225676A2 - Method and server for constructing knowledge base

Info

Publication number: EP2225676A2
Application number: EP08861527A
Authority: EP
Inventors: Sung Pil Choi; Jerry Hyeon Seo; Jin Suk Kim; Yun Soo Choi; Hwa Mook Yoon; Sun Hwa Han
Original assignee: Korea Institute of Science and Technology KIST; Korea Institute of Science and Technology Information KISTI
Current assignee: Korea Institute of Science and Technology KIST; Korea Institute of Science and Technology Information KISTI
Priority date: 2007-12-18
Filing date: 2008-12-16
Publication date: 2010-09-08
Also published as: KR20090065625A; WO2009078649A2; EP2225676A4; WO2009078649A3; KR100938830B1

Abstract

The present invention relates to a method and server for constructing a knowledge base. According to the method, a schema item is defined. When resource information, including at least one of a file, a difficulty and an arrival goal, is input, identifiers unique to the resource information are created, and the created identifiers are assigned to the resource information. Pieces of attribute information are extracted from the input resource information by analyzing the resource information. An associated relationship between the respective pieces of extracted attribute information is set based on the defined schema item. The set associated relationship is stored. Accordingly, difficulty information about specific resources of field experts, knowledge about a research sequence with respect to lower element fields of a current field, and so on can be directly expressed in the knowledge base.

Description

METHOD AND SERVER FOR CONSTRUCTING KNOWLEDGE

BASE

Technical Field

[1] The present invention relates to a method and server for constructing a knowledge base, wherein, when resource information is input, pieces of attribute information are extracted from the resource information by analyzing the resource information, and the associated relationship of the respective pieces of the extracted attribute information is set and stored based on a previously defined schema item.

[2]

Background Art

[3] The publication volume of scientific and technological documents is gradually increasing, and information, which can be obtained through an Internet, has been excessively increased. Accordingly, such excessive information serves as a factor, which hinders correct decision, to people who must understand a tendency of research through the information, secure competitiveness between companies on the basis of the information, set the direction of research and development according to the general trend, and be prepared for a future challenge.

[4] Traditional information analysis method in which an information user obtains and analyzes information on the basis of a personal knowledge and experience from all information sources is disadvantageous in that much time is taken for an expert to obtain and analyze the information, biased information can be collected and information can be analyzed on a biased view according to an expert's point of view, and so on.

[5] As one method of improving the disadvantages, research is doing on the development of an information analysis system, which performs some of tasks, performed by a person, using a computer.

[6] Further, a lot of rising researchers who try to start new research into specific research fields experience lots of difficulties in finding the start points of their researches.

[7] If someone has a good luck, seniors or professor who try to help and support the person actively and devotedly may become a solution to the difficulties on the start point, but persons have to solve the difficulties by themselves.

[8] The start point for research activities in a new field may be several, but, representatively, may include general articles and description sentences drafted by the experts of specific fields, basic personal information about experts who output unique research results in specialty fields or are doing vigorous activities, textbook information of special field, and the like. [9] Gathering of information, that is, a preparation task for performing research in earnest is considered as a task, which is very difficult and takes much time, even in the present time of an advanced Internet.

[10] In contrast to the difficulty, with respect to experts who are already experienced in specific fields, such information is close to deserved axiom. Accordingly, what the field experts handle this basic level of information may result in losing their time in some ways.

[11] However, many efforts have been made for a long time in order to obtain the first step of solving the problem in which how beginner researcher can obtain such information easily using a methodology, such as knowledge expression and knowledge base construction in artificial intelligence fields. A field-based expert system, which utilizes a knowledge base constructed by an elaborately defined knowledge expression scheme, has been developed a lot.

[12] A mobile computer having an in-depth knowledge on a specific field, if any, may be considered as a system including one specialty field knowledge about the field, which covers all pieces of knowledge for some centuries. This system may also be very useful to administrators who must perform situation assessment or long-range planning.

[13] According to another opinion about the definition of an expert system, the expert system is considered as one large computer program from a viewpoint of the computer science. In other words, an expert system or a knowledge-based system is defined as one computer program including subject- specific knowledge of a number of experts.

[14] These kinds of programs were first developed in the 1960s and 1970s and began being utilized commercially in the 1980s. The most general type of an expert system has the following characteristics.

[15] First, the expert system includes a regular set, which is necessary to analyze a detailed level of information about a problem that a user wants to solve.

[16] Second, the expert system performs problem analysis using a variety of mathematical methodologies on the basis of analyzed information.

[17] Finally, the expert system provides a user action scenario, which is necessary to solve problems on the basis of analyzed results or modify error of the expert system.

[18] Accordingly, a user can make clear his ultimate goal through an interaction with the expert system and may be helped to modify erroneous results of the system. In this information exchange process, the expert system utilizes an inference support.

[19] As described above, in order to make and substantially utilize one expert system, a target domain must be defined clearly and a field thereof must be expressed sufficiently.

[20] In other words, if a target domain is too wide, there are difficulties in the knowledge expression stage that is manually performed. This indispensably accompanies a representative power short phenomenon. The knowledge base constructed as above has very low inference ability and is difficult to obtain a good grade in terms of a feeling of satisfaction or usage by a user.

[21] It can be seen that there is a significant gap between a logic -based knowledge base construction, which is researched in the traditional artificial intelligence field, and standard framework-based knowledge expression (RDF, RDFS, OWL, etc.), which is described in a semantic web and somewhat simplified and intuitive.

[22] Further, this gap offers lots of points of discussion between scholars who are doing research on the two fields. The main subject of the dispute is related to inference. It can be seen that these points of discussion are important factors when viewing Jena, that is, a Java-based semantic web framework that is most widely used.

[23] In other words, inference support by Jena is mostly limited to the handling of a transitive relation or an entailment statement. Inference support by Jena mainly omits a first-order logic, which is handled in the traditional artificial intelligence field, or a description logic-based general-purpose inference support. This is because the present OWL is based on a description logic, and a usage scenario or a standard framework about OWL-DL, which covers the greater part of the description logic has not yet been clearly defined.

[24] Accordingly, one of the most important objects to construct the knowledge base is inference.

[25] Inference from a standpoint of traditional artificial intelligence is an instance set, which is expressed on the basis of a basic logic, such as a description logic and a first- order logic, that is, a task of finding out new knowledge from a knowledge base. In order for this complicated type of inference to be performed, a target instance set must be expressed very elaborately and must not include error.

[26] However, field-based experts are persons who have specialty knowledge and experience about their fields as it says. There are disadvantages in that those persons become familiar with a knowledge expression method based on the logic of artificial intelligence and express their knowledge using the knowledge expression method.

[27] Further, conventional knowledge base is disadvantageous in that results do not give a feeling of satisfaction to users and are not properly utilized because it is complicated and has inefficient inference support.

[28]

Disclosure of Invention Technical Problem

[29] Accordingly, the present invention has been made in view of the above problems occurring in the prior art, and an object of the present invention is to provide a knowledge base construction method and server, which constructs an intuitive knowledge base on the basis of a user's current knowledge level and target level in order to reduce the difficulties of coming researchers, which occur at the beginning stage of researches, by utilizing experts who are placed in a certain level in specific fields.

[30] Another object of the present invention is to provide a knowledge base construction method and server, which is capable of providing a knowledge base that forgives a complicated and inefficient inference support in traditional artificial intelligence and is intuitive and easy to construct.

[31] Still another object of the present invention is to provide a knowledge base construction method and server, which is capable of expressing difficulty information about the specific resources of field experts, knowledge about a research sequence with respect to lower element fields of a current field, and so on in the knowledge base.

[32] Further still another object of the present invention is to provide a knowledge base construction method and server, which helps field experts to easily express their expert knowledge and allowing a user to obtain the most essential information necessary to perform researches.

[33]

Technical Solution

[34] To achieve the above objects, according to an aspect of the present invention, there is provided a method of constructing a knowledge base, including the steps of (a) defining a schema item, (b) when resource information, including at least one of a file, a difficulty and an arrival goal, is input, creating identifiers unique to the resource information and assigning the created identifiers to the resource information, (c) extracting pieces of attribute information from the input resource information by analyzing the resource information, and (d) setting an associated relationship between the respective pieces of extracted attribute information based on the defined schema item and storing the set associated relationship.

[35] The schema item is a RDF-based schema item, and the schema item includes a mother class classifying resources, a child class, that is, a detailed type of each mother class, and attribute list information included in each child class.

[36] The file includes at least one of articles, books, and web documents. The difficulty is a difficulty with respect to the file and includes one of easy, medium, and difficult. The arrival goal is an ultimate goal in a specific field, and includes one of skin-deep, basic, and advanced.

[37] The step (b) includes the steps of when a knowledge base construction command is input, displaying a resource information input screen, and, when resource information, including a file, a difficulty, and an arrival goal, is input through the resource information input screen, creating identifiers unique to the resource information and assigning the created identifiers to the resource information.

[38] The step (c) includes extracting attributes, which correspond to attribute list information previously defined based on the schema item, and values of the attributes by analyzing the resource information.

[39] The step (d) includes setting an associated relationship with corresponding child classes based on the defined schema item with respect to the respective pieces of extracted attribute information and setting an associated relationship with mother classes associated with the child classes.

[40] According to another aspect of the present invention, there is provided a knowledge base construction server for constructing a knowledge base, including a resource information receiving unit for receiving resource information, including a file, a difficulty, and an arrival goal, a resource identifier generating unit for, when the resource information is input through the resource information receiving unit, creating identifiers unique to the resource information and assigning the created identifiers to the resource information, an attribute information extraction unit for extracting pieces of attribute information, including attributes corresponding to an attribute list defined in a previously defined schema item and values of the attributes, from the resource information received from the resource information receiving unit by analyzing the resource information, and an associated relationship setting unit for setting an associated relationship between the respective pieces of attribute information extracted from the attribute information extraction unit based on the previously defined schema item, mapping the attribute information to the identifiers created in the resource identifier generating unit, and storing the mapping attribute information.

[41] The associated relationship setting unit sets an associated relationship with corresponding child classes with respect to each piece of attribute information extracted from the attribute information extraction unit and sets an associated relationship with a mother class associated with the child classes.

[42]

Advantageous Effects

[43] Accordingly, the present invention may provide a knowledge base construction method and server, which is capable of constructing an intuitive knowledge base on the basis of a user's current knowledge level and target level in order to reduce the difficulties of coming researchers, which are generated at the beginning stage of researches, by utilizing experts who are placed in a certain level in specific fields. [44] Further, the present invention may provide a knowledge base construction method and server, which is capable of expressing difficulty information about the specific resources of field experts, knowledge about a research sequence with respect to lower element fields of a current field, and so on in the knowledge base.

[45] Further, the present invention may provide a knowledge base construction method and server, which helps field experts to easily express their expert knowledge and allowing a user to obtain the most essential information necessary to perform researches.

[46]

Brief Description of the Drawings

[47] Further objects and advantages of the invention can be more fully understood from the following detailed description taken in conjunction with the accompanying drawings in which:

[48] FIG. 1 is a diagram showing the configuration of a knowledge base construction system according to the present invention;

[49] FIG. 2 is a block diagram schematically showing the configuration of a knowledge base construction server according to the present invention;

[50] FIG. 3 is a flowchart showing a method of constructing a knowledge base according to the present invention;

[51] FIG. 4 is a diagram showing a RDF schema item according to the present invention;

[52] FIG. 5 is a flowchart showing a method of constructing a knowledge base on the basis of a RDF schema item according to the present invention; and

[53] FIG. 6 illustrates a resource information input screen according to the present invention.

[54] <Description of reference numerals of principal elements in the drawings>

[55] 100: client 110: knowledge base construction server

[56] 210: resource information receiving unit

[57] 220: resource identifier generating unit

[58] 230: attribute information extraction unit

[59] 240: associated relationship setting unit

[60] 250: knowledge base

Mode for the Invention

[61] Hereinafter, detailed contents of the above-described objects, technical configurations, and operational effects thereof according to the present invention will be more clearly understood from the detailed description of the present invention with reference to the attached drawings.

[62] FIG. 1 is a diagram showing the configuration of a knowledge base construction system according to the present invention.

[63] Referring to FIG. 1, a knowledge base construction system includes a client 100 in which knowledge experts input resource information, and a knowledge base construction server 110 for constructing a knowledge base using resource information received from the client 100.

[64] The client 100 may include a wired communication terminal, a wireless communication terminal, etc.

[65] If resource information, including a file, difficulty, an arrival goal, etc., is input by a knowledge expert, the client 100 transmits the input resource information to the knowledge base construction server 110 over a communication network.

[66] If the resource information is received from the client 100, the knowledge base construction server 110 analyzes the resource information, extracts pieces of attribute information from the resource information based on a previously defined RDF schema item, sets a associated relationship between the respective pieces of extracted attribute information, and constructs a knowledge base.

[67] Here, the knowledge base includes essential information, which is necessary for new researcher to start researches in specific fields.

[68] Before the knowledge base is constructed, the entire constituent elements are formalized, that is, RDF schema items are defined in the knowledge base construction server 110.

[69] That is, the knowledge base may be expressed in a set of elements which belong to a virtual set as in Equation 1.

[70] [Equation 1]

[71]

[72] Here, is the resource set constituting a specific expert field, and comprises a resource identifier I, a associated relationship C, a resource difficulty D, a user arrival goal G, and a target entity O.

[73] The resource identifier set I is a set of individual identifiers, which is able to distinguish all resources existing within a specific knowledge base. A method of expressing elements of the set may be very various. For example, RDF, that is, the core framework of a semantic web adopts a Uniform Resource Identifier (URI) as a method of expressing RDF.

[74] The associated relationship set C includes pieces of relation information, which connect two entities, as its elements.

[75] The resource difficulty is a content level of current resources which can be determined by field experts. For example, when considering books about specific fields, it can be seen that books having titles, such as Introduction to -' and 'Elementary -', belong to resources with a low difficulty although they belong to the same field. [76] However, it is assumed that books containing the recent theory or idea in the corresponding field or books determined to be an upper level in the same field have a high difficulty.

[77] The user arrival goal G has the same concept and characteristic as those of the resource difficulty. The user arrival goal G differs from the resource difficulty in that it is used as a difficulty for which a user (rising researcher) now wants to reach in his selected field. The resource difficulty D and the user arrival goal G may be expressed as follows.

[78]

[79]

[80] Here, the numerals within the parentheses are identifiers, denoting respective constitution items.

[81] A field expert who is responsible for constructing a knowledge base designates a resource difficulty based on his determination every element resource. Further, a user who utilizes a constructed knowledge base may search for resources to which reference must be currently made indispensably by inputting his arrival goal level.

[82] Finally, the target entity set O is a set, including objective resources connected to resource identifiers through relation information as its elements, and is expressed in the following Equation 2.

[83] [Equation 2]

[84]

[85] Here, is a set of character strings consisting of a natural language that is able to express a specific entity. In other words, the target entity set may be considered as the union of an identifier set, which can be used to identify and indicate resources, and the literary language set. This is for the purpose of, when a number of entities exist in a knowledge base, defining a target entity set so that property description about a specific entity is possible when the specific entity and other entity are connected through associated relationship information.

[86] The associated relationship between the above-described individual entities may be expressed by a function R. The function receives the resource identifier I, the associated relationship C, the resource difficulty D, and the user arrival goal G, and outputs a corresponding target entity O as in Equation 3.

[87] [Equation 3]

[88] R:I*C*D*G -> O

[89] The knowledge base construction server 110 that plays the above-described role is described in detail with reference to FIG. 2.

[90] FIG. 2 is a block diagram schematically showing the configuration of a knowledge base construction server according to the present invention. [91] Referring to FIG. 2, the knowledge base construction server includes a resource information receiving unit 210, a resource identifier generating unit 220, an attribute information extraction unit 230, a associated relationship setting unit 240, and a knowledge base 250.

[92] The resource information receiving unit 210 functions to receive resource information, including a file, difficulty and an arrival goal, from a user. Here, the file refers to a web page, some of a web page, books, articles, and so on.

[93] The resource identifier generating unit 220 creates identifiers unique to the resource information and assigns the unique identifiers to the resource information, when the resource information is received through the resource information receiving unit 210. Here, the identifiers may be, for example, an URI. That is, a unique URI is attached to each resource item, and pieces of characteristic information are expressed in attribute form, so all resources are identified through URIs. Accordingly, the resource identifier generating unit 220 generates unique URIs with respect to the resource information.

[94] The attribute information extraction unit 230 extracts attributes and values matching the respective extracted attributes from the resource information by analyzing the resource information.

[95] That is, the attribute information extraction unit 230 extracts attributes, belonging to an attribute list defined in a previously defined RDF schema item, and values of the attributes by analyzing the resource information.

[96] Each RDF schema item comprises mother classes classifying resources, child classes, that is, the detailed types of mother classes, and attribute list information included in each child class. Accordingly, the attribute information extraction unit 230 extracts attributes, belonging to the attribute list information, and values of the extracted attributes by analyzing the resource information.

[97] For example, if input resource information includes an article and difficulty, and an attribute list, which is included in a child class defined in a RDF schema item, is pub_Title, pub_Author, pub_Tech, pub_Year, difficulty_Level, person_HomePage, person_Name, tech_Area, and download_Link, the attribute information extraction unit 230 extracts attributes of pub_Title, pub_Author, pub_Tech, pub_Year, difficulty_Level, download_Link and values of the extracted attributes from the resource information.

[98] Thereafter, the attribute information extraction unit 230 transmits attribute information, including the extracted attributes and the values of the attributes, to the associated relationship setting unit 240.

[99] The associated relationship setting unit 240 sets an associated relationship between the respective pieces of attribute information, extracted from the attribute information extraction unit 230, based on the previously defined RDF schema item, and stores the set associated relationship. Accordingly, a knowledge base is constructed.

[100] That is, the associated relationship setting unit 240 sets child classes, corresponding to the respective pieces of attribute information extracted from the attribute information extraction unit 230, and an associated relationship therebetween by determining the child classes, sets mother classes, including the child classes, and an associated relationship therebetween, and stores the set mother classes and the set associated relationship in the knowledge base 250.

[101] FIG. 3 is a flowchart showing a method of constructing a knowledge base according to the present invention. FIG. 4 is a diagram showing a RDF schema item according to the present invention.

[102] Referring to FIG. 3, the knowledge base construction server first defines a RDF schema item in order to construct a knowledge base (S300).

[103] The RDF schema item refers to a RDF instance set, that is, an element item necessary to express the knowledge base, and is described below with reference to FIG. 4.

[104] Referring to FIG. 4, each RDF schema item comprises a total of three mother classes, including 'Person', 'Publication', and 'Technology' on a type basis. Child classes, indicating detailed types, exist in each mother class. Further, each child class includes attribute list information, indicating an instance to which each class belongs.

[105] For example, a class 'Book', that is, a child class of the class 'Publication' includes attribute sets, such as 'pub_Title' corresponding to a book title and 'pub_Author' corresponding to a writer. The attributes can be seen as the elements of an associated relationship set.

[106] The classes 'Publication' and 'Paper' also include an associated relationship such as a mother-child relationship, as well as the associated relationship information that is directly seen.

[107] Of the attribute sets, 'difficulty_Level' is a slot that is able to express difficulty information of current resources. A field expert who constructs a knowledge base determines the difficulty of a target resource that will be constructed and marks a value in the slot.

[108] When the resource information, including a file and difficulty, is input after the RDF schema item is defined (S302) as described above, the knowledge base construction server extracts pieces of attribute information based on the defined RDF schema item by analyzing the resource information and sets an associated relationship between the respective pieces of extracted attribute information (S304).

[109] A method in which the knowledge base construction server extracts attribute information based on the RDF schema item and setting an associated relationship with respect to the attribute information will be described in detail below with reference to FIG. 5. [110] If the step S304 is performed, the knowledge base construction server stores information whose associated relationship has been set, thereby constructing a knowledge base (S306).

[I l l] FIG. 5 is a flowchart showing a method of constructing a knowledge base based on a RDF schema item according to the present invention. FIG. 6 illustrates a resource information input screen according to the present invention.

[112] Referring to FIG. 5, the knowledge base construction server receives resource information (S500), and creates a URI unique to the resource information and assigns the created URI to the resource information (S502).

[113] That is, if a user inputs a knowledge base construction command in order to construct a knowledge base with respect to articles, the knowledge base construction server provides a resource information input screen 600 as shown in FIG. 6.

[114] The resource information input screen 600 includes a file input area 610, a difficulty input area 620, and an arrival goal input area 630.

[115] The user may input web pages, books, articles, etc. in the file input area 610. If a corresponding file is a web page, the user may input a web page address, and, if a corresponding file is an article, the user may fetch and store the article.

[116] The difficulty input area 620 is an area in which the difficulty of a corresponding file is input. One of difficulties 'easy', 'medium', and 'difficult' may be input in the difficulty input area 620.

[117] The arrival goal input area 630 is an area in which the ultimate goal to be reached in a specific field is input. In this area, one of resource difficulty levels 'skin-deep', 'basic', and 'advanced' may be input. The 'skin-deep' requires only resources corresponding to the resource difficulty level 'easy'. The 'basic' requires only resources corresponding to the resource difficulty levels 'easy' and 'medium'. The 'advanced' requires only resources corresponding to all the resource difficulty levels 'easy', 'medium', and 'difficult'.

[118] After the resource information is input through the resource information input screen 600 as described above, the knowledge base construction server creates and assigns unique URIs for identifying the resource information.

[119] Thereafter, the knowledge base construction server extracts attributes by analyzing the resource information (S504). That is, since an attribute list has already been defined in a RDF schema items, the knowledge base construction server extracts attributes corresponding to the attribute list, which is defined in the RDF schema item, by analyzing the resource information.

[120] Thereafter, the knowledge base construction server extracts values corresponding to the extracted attributes (S506).

[121] For example, in the case in which a file included in the resource information is an article, the knowledge base construction server extracts person_HomePage, person_Name, and tech_Area with respect to a writer of the article by analyzing the article, and extracts pub_Title, pub_Author, pub_Tech, pub_Year, etc. from the contents of the article and an attribute, including difficulty and an arrival goal, and a value thereof from the resource information.

[122] Thereafter, the knowledge base construction server sets an associated relationship between respective attributes based on the previously defined RDF schema item (S508).

[123] That is, each RDF schema item comprises mother classes classifying resources, child classes, that is, detailed types of each mother class, and an attribute list included in each child class. Accordingly, the knowledge base construction server sets an associated relationship with a corresponding child class with respect to each of the extracted attributes and sets an associated relationship with a mother class associated with the child class.

[124] For example, in the case in which a RDF schema item is defined as shown in FIG. 4, the knowledge base construction server connects each attribute to a child class and connects the child class to a mother class associated with the child class. That is, a class 'Book', that is, a child class of the class 'Publication' has attribute sets, such as 'pub_Title' corresponding to a book title and 'pub_Author' corresponding to a writer.

[125] If the associated relationships are set as in the step S508, the knowledge base construction server maps information whose associated relationships have been set to the created URIs and stores the results (S510). Accordingly, a knowledge base is constructed.

[126] Those skilled in the art will understand that the invention can be implemented in detailed forms without changing the technical spirit or indispensable characteristics of the present invention. Therefore, the above-described embodiments should be construed to be illustrative and limitative from all aspects. Furthermore, the scope of the present invention is defined by the appended claims rather than the above detailed description. Thus, the present invention should be construed to cover all modifications or variations induced from the meaning and range of the appended claims and their equivalents.

Claims

[1] A method of constructing a knowledge base, the method comprising the steps of:

(a) defining a schema item;

(b) when resource information, including at least one of a file, a difficulty and an arrival goal, is input, creating identifiers unique to the resource information and assigning the created identifiers to the resource information;

(c) extracting pieces of attribute information from the input resource information by analyzing the resource information; and

(d) setting an associated relationship between the respective pieces of extracted attribute information based on the defined schema item and storing the set associated relationship.

[2] The method as claimed in claim 1, wherein the schema item is a RDF-based schema item.

[3] The method as claimed in claim 1, wherein the schema item includes a mother class classifying resources, a child class, that is, a detailed type of each mother class, and attribute list information included in each child class.

[4] The method as claimed in claim 1, wherein the file includes at least one of articles, books, and web documents.

[5] The method as claimed in claim 1, wherein the difficulty is a difficulty with respect to the file, and includes one of easy, medium, and difficult.

[6] The method as claimed in claim 1, wherein the arrival goal is an ultimate goal in a specific field, and includes one of skin-deep, basic, and advanced.

[7] The method as claimed in claim 1, wherein the step (b) comprises the steps of: when a knowledge base construction command is input, displaying a resource information input screen; and when resource information, including a file, a difficulty, and an arrival goal, is input through the resource information input screen, creating identifiers unique to the resource information and assigning the created identifiers to the resource information.

[8] The method as claimed in claim 1, wherein the step (c) includes extracting attributes, which correspond to attribute list information previously defined based on the schema item, and values of the attributes by analyzing the resource information.

[9] The method as claimed in claim 1, wherein the step (d) includes setting an associated relationship with corresponding child classes based on the defined schema item with respect to the respective pieces of extracted attribute information and setting an associated relationship with mother classes associated with the child classes.

[10] A knowledge base construction server for constructing a knowledge base, comprising: a resource information receiving unit for receiving resource information, including a file, a difficulty, and an arrival goal; a resource identifier generating unit for, when the resource information is input through the resource information receiving unit, creating identifiers unique to the resource information and assigning the created identifiers to the resource information; an attribute information extraction unit for extracting pieces of attribute information, including attributes corresponding to an attribute list defined in a previously defined schema item and values of the attributes, from the resource information received from the resource information receiving unit by analyzing the resource information; and an associated relationship setting unit for setting an associated relationship between the respective pieces of attribute information extracted from the attribute information extraction unit based on the previously defined schema item, mapping the attribute information to the identifiers created in the resource identifier generating unit, and storing the mapping attribute information.

[11] The knowledge base construction server as claimed in claim 10, wherein the schema item is a RDF-based schema item.

[12] The knowledge base construction server as claimed in claim 10, wherein the schema item includes a mother class classifying resources, a child class, that is, a detailed type of each mother class, and attribute list information included in each child class.

[13] The knowledge base construction server as claimed in claim 10, wherein the associated relationship setting unit sets an associated relationship with corresponding child classes with respect to each piece of attribute information extracted from the attribute information extraction unit and sets an associated relationship with a mother class associated with the child classes.