JP2001306603A - Structured database generating device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To solve such a problem that a conventional structured database generating device requires a word database wherein a sufficient number of words are previously stored and, particularly, a large-scale data base requires considerable operation for word registration. SOLUTION: This structured database generating device is equipped with a name data storage means 3, a node name candidate extracting means 4 which makes the longest-length matching comparison between character strings of pieces of name data from the beginning to the end to extract a node name candidate, a node name selecting means 5 which selects a proper node name, a hierarchical structure generating means 6 which generates a structure database by constituting a hierarchical structure of selected node names, etc., and a hierarchical structure storage means 7 which stores the structured database.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a structure for extracting constituent elements by hierarchy from structural information included in input name data and defining a mutual relationship between the constituent elements to create a structured database. The present invention relates to a structured database creation device.

[0002]

2. Description of the Related Art In order to distinguish an object managed in a business system or the like from other objects, a name expressing the object by connecting the component names of respective layers relating to the structure of the object. Often given as. A typical example of using such a name expression is facility data of a plant. As the name data used in the facility management of the plant, for example, the expression “XX tank XX pump / valve” is used. In other words, although the target of the data is a valve called “△△ valve”, in order to distinguish it from other valves, [［pump attached to “xx pump” provided in “「 tank ” The name is added with structural information such as "valve". Therefore, by extracting the component information of each layer from the structure information of the object included in the name data, it is possible to create a structured database that reflects the structure of the equipment to be processed.

[0003] The structured database can be easily obtained from a mere set of name data because the structure of the object represented by the components related to the object and the mutual relation between the components can be easily grasped. It has the following advantages that cannot be obtained. First, an efficient search can be provided. For example, a search related to the interrelation between components, such as a search for the entire facility attached to a certain facility,
Only by following the link related to the mutual relationship (in this case, the link indicating the affiliation relationship), it is possible to execute the process at high speed without any excess or shortage.

[0004] Second, maintenance of a database is facilitated. When digitizing an existing large-scale database, the management number system, naming rules, and the like are often reviewed. In such a case, if a structured database expressing the constituent elements of the object and the relationship between the elements is created, the entire structure of the object can be easily grasped. Rules concerning number assignment, name naming, and the like can be efficiently determined from the top down.

Third, database integration is facilitated. For example, in the case of equipment data, an attempt is made to integrate a database of similar plants and manage a plurality of equipments in an integrated manner in order to improve the efficiency of equipment management work.
At this time, if a structured database is created for each facility, it is possible to easily determine the integration of the databases with respect to identification of a part that can be structurally shared and a part that is maintained as local data.

For the above reasons, a method of creating a structured database from input name data has been proposed, particularly in connection with the construction of an equipment management system. FIG.
FIG. 1 is a diagram showing a configuration of a conventional structured database creation device described in, for example, Japanese Patent Application Laid-Open No. 6-50778. In FIG. 14, 101a and 101b are name data input by the user, 102 is the structure of the plant equipment,
A word database for storing names relating to functions and failure modes for each category in advance, 103 is a preventive maintenance database as a structured database for storing preventive maintenance information, and 104 is input name data 101a, 101b.
This is an information processing means for performing a division process with respect to the word database 102 and storing the divided data structurally in the preventive maintenance database. Note that the word database 102 includes
A thesaurus in which synonyms and synonyms are collected is prepared. By referring to this thesaurus, it is possible to recognize that “plant A” and “plant A ′” represent the same object, for example.

Next, the operation will be described. The information processing unit 104 performs a search on the name data 101a and 101b input by the user with reference to the word database 102, and uses the matched words to search the name data 101a and 101b.
a, 101 b are divided, converted into a hierarchical structure, and stored in the preventive maintenance database 103.

For example, in the word database 102, "nuclear power plant demo
Plant, system-level word data "reactor recirculation system", equipment-level word data "pump"
By storing “casing” in advance as word data at the part level, the input data “nuclear power plant d
emo plant reactor recirculation pump casing "
It can be converted and stored in a hierarchical structure of "nuclear power plant demo plant"-"reactor recirculation system"-"pump"-"casing". In addition, the thesaurus in the word database 102 includes “reactor recirculation system” and “PL
If "R system" is registered as having the same meaning, the input data "nuclear power plant demo plant PLR
The "system pump casing" can be equivalently converted into a hierarchical structure of "nuclear power plant demo plant"-"reactor recirculation system"-"pump"-"casing" and accumulated.

[0009]

Since the conventional structured database creating apparatus is configured as described above, it is necessary to register a sufficient amount of words in the word database 102 in advance, or to input name data. At the same time, the user needs to interactively add words to the word database 102. However, in order to create the word database 102 in advance, it is necessary to foresee all the name data that may be input, manually investigate each word constituting the name data, and register them in advance. In particular, there is a problem that a large amount of work occurs in a large-scale database.

On the other hand, when words are added interactively in parallel with the input of the name data, it is necessary to input the word data one by one. There was a problem of becoming.

[0011] In addition, since the names of facilities generally differ from plant to plant in many cases, the above-mentioned word database creation work must be performed for each plant, and there is a problem that more work is required. Was.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-described problems, and it is not necessary to prepare a word database in advance, and a component is extracted for each layer from the structure information included in the name data. It is an object of the present invention to obtain a structured database creation device for automatically creating a structured database.

[0013]

According to the present invention, there is provided a structured database creating apparatus comprising: a name data storage unit for storing input name data; nodes corresponding to constituent elements arranged hierarchically; A hierarchical structure storing means for storing a structured database configured by defining the interrelationship between them, and reading the name data stored in the name data storing means, and storing all other name data for each name data. A longest match comparison from the beginning and a longest match comparison from the end are performed on the character string forming the name data between the two, and a node name candidate given as a character string representing a node candidate constituting the structured database is obtained. The node name candidate extracting means to be extracted and the node name candidates extracted by the node name candidate extracting means are presented, and from these node name candidates, Node name selection means for allowing a user to select a node name of an appropriate node as a component constituting the structured database, and a node having the node name selected by the node name selection means and name data stored in name data storage means And a hierarchical structure creating means for creating a structured database by defining the interrelationship between the nodes in a hierarchical manner.

The structured database creating apparatus according to the present invention is provided with a hierarchical structure editing means for enabling a user to edit the structured database stored in the hierarchical structure storage means.

[0015] A structured database creating apparatus according to the present invention comprises a synonym definition means for enabling a user to define a synonym, and a synonym information storage means for storing synonyms defined using the synonym definition means. And synonym conversion means for referring to the synonym information stored in the synonym information storage means, and converting the name data and the node name into representative terms having the same meaning.

An apparatus for creating a structured database according to the present invention reads an extraction auxiliary rule storing means for storing an extraction auxiliary rule applied when extracting a node name candidate, and reads name data stored in a name data storage means. In each of the name data, the longest match comparison from the beginning and the longest match comparison from the end with respect to the character string forming the name data with all other name data are performed, and the nodes constituting the structured database A node name candidate given as a character string representing the candidate of the extracted node name candidate, and if the extraction process of the obtained node name candidate corresponds to any of the rules stored in the extraction auxiliary rule storage means, Node name candidate extracting means for discarding name candidates.

A structured database creation device according to the present invention includes a name correction rule storage unit for storing a name correction rule defining a correction method for a term included in name data, and a name correction rule storage unit for storing name data constituting input data. A name correction unit that applies a name correction rule stored in a name correction rule storage unit and accumulates name data converted based on the name correction rule in the name data storage unit. .

[0018]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described below. Embodiment 1 FIG. FIG. 1 is a diagram showing a configuration of a structured database creation device according to Embodiment 1 of the present invention. 1
Is input data composed of one or a plurality of name data, 2 is a user, 3 is name data storage means for storing input name data, and 4 is a configuration which tests each name data stored in the name data storage means 3. A node name candidate extracting means for extracting a character string representing a node candidate corresponding to an element from the name data, 5 discloses the node name obtained by the node name candidate extracting means 4 to the user 2, and configures the components of each node. Node name selecting means for allowing the user 2 to select whether or not the name is appropriate, 6 is a hierarchical structure creating means for creating a structured database based on name data and an appropriate node name selected for the name data, 7 is a hierarchical structure Hierarchical structure storage means for storing the structured database created by the creation means 6,
8 is name data storage means 3, node name candidate extraction means 4,
This is a structured database creation device including a node name selection means 5, a hierarchy structure creation means 6, and a hierarchy structure storage means 7.

FIG. 2 is a diagram showing an example of input data composed of a plurality of name data. In FIG. 2, L1 to L1
0 indicates a facility name in the power plant. Regarding the input method of the input data, for example, the input person may interactively input information described on a paper form, or collectively transmit information stored in an existing database using a program. May be entered. In any case, the name data is stored in the name data storage means 3 while retaining the format shown in FIG.

FIG. 3 is a diagram showing a hierarchical structure in the structured database. The hierarchical structure shown in FIG. 3 is created by dividing each name data of the input data 1 shown in FIG. 2 into node names. As shown in FIG.
The hierarchical structure is composed of components called nodes indicated by square frames and lines (links) expressing various relationships between these nodes.

In FIG. 3, reference numeral 11 denotes a class node representing a conceptual object, and reference numeral 12 denotes an instance node representing an actual physical object. The name data L1 to L10 shown in FIG. 2 will eventually form a hierarchical structure as this instance node. Reference numeral 13 denotes an is_a link indicating a mutual relationship between the class nodes 11, and expresses, for example, a relationship [“the inlet valve” is a “valve”]. In FIG. 3, the node on the left side of the is_a link has a higher degree of abstraction than the node on the right side. 14
Is the root node which is located on the leftmost side and is the most abstract virtual node among all the nodes,
This is a special node prepared by the structured database as an initial value. Reference numeral 15 denotes an inst representing a "concept-entity" relationship between the class node 11 and the instance node 12.
ance_of link. Reference numeral 16 denotes a has_a link indicating the ownership relationship between the instance nodes 12, and represents, for example, a relationship "[A condenser] has" A condenser vacuum pump "]. Also, has_a link 16
The arrow means that the node on the proximal end side of the arrow owns the node on the distal end side of the arrow.

When there are a plurality of instances of the same type, such as "A condenser" and "B condenser", there is a class "condenser" as a set of these instances. However, if there is only one instance of the same type as the "condensed water recovery tank", such as the "condensed water recovery tank", the class "condensed water recovery tank" is not created.

Next, the flow of the entire structured database creation process will be described. FIG. 4 is a flowchart showing a structured database creation method. First, the name data storage means 3 reads data interactively from a user or collectively reads name data from an existing database and accumulates the input name data (step ST1). Next, the node name candidate extraction means 4 reads the data stored in the name data storage means 3 and extracts a character string representing a class node candidate in the hierarchical structure from the name data (step ST2). Next, the node name selecting means 5 presents the character strings extracted by the node name candidate extracting means 4 to the user in ascending order of length, and asks the user whether each character string is appropriate as a class node name. The user is made to make a determination, and the result of the determination is input using a mouse or the like (step ST3).

Next, the hierarchical structure creating means 6 executes step S
At T3, the class node 11 is generated based on the class node name determined to be appropriate by the user, and a hierarchical structure is created by automatically establishing an is_a link between the class nodes. The data is stored in the storage means 7 (step ST4). Note that the process relating to step ST4 may be executed each time a node name is selected by the user in step ST3.
It may be executed in a batch process after all the node names are selected in step ST3. However, in each case, the selected class node names are input in ascending order of the character string length.

In step ST4, when the creation of the hierarchical structure related to the class nodes is completed, the hierarchical structure creating means 6
Reads the name data stored in the name data storage means 3 in ascending order of length, and reads the instance node 12
Is generated and added to the hierarchical structure (step ST5). At this time, an instance_of link is established between the class node 11 of the hierarchical structure stored in the hierarchical structure storage means 7 and a has
The creation of the structured database is completed by linking the __a link.

Next, a description will be given of a node name candidate extraction process executed in step ST2. FIG. 5 is a flowchart illustrating the processing steps of the node name candidate extraction processing.
First, a node name candidate list NL is created and the node name candidate list NL is made an empty list (step ST1).
1). Next, using the counter variable i, the subsequent processing is repeatedly executed by the number of all name data (step ST1).
2). Then, in this repetition processing, first, the i-th name data Di is extracted (step ST13).
Next, the local variables Lf, Lb, SL, and S are initialized (step ST14). Note that Lf and Lb are the number of characters,
SL is a character string list, and S is a variable for storing a character string, and is initialized to 0, an empty list, and an empty character string, respectively.

Next, using the counter variable j, the subsequent processing is repeatedly executed by the number of all name data (step S).
T15). In this repetition processing, first, if i and j are the same, the process immediately returns to step ST15 so as not to compare the same name data (step ST15).
16). If i and j are different, the j-th name data Dj is extracted (step ST17). Then, the following two character string comparison processes are performed between the name data Di and the name data Dj.

As a first character string comparison process, first, the character string of the name data Di and the character string of the name data Dj are compared from the beginning, and the number of characters that are continuously matched from the beginning is L.
f ′ and the remaining character strings obtained by removing the matched character strings from the name data Di and Dj are set to Si and Sj, respectively (step ST18). Next, the number of characters Lf 'is compared with Lf given as the number of characters that have longest matched up to that point in the first character string comparison process for the name data Di (hereinafter, referred to as the first longest matching character number) ( Step ST
19). In the comparison in step ST19, L
If f ′ = Lf, that is, if the current number of matching characters is equal to the first longest matching character number, the remaining character strings Si, S
j is added to the character string list SL. Step ST
In the comparison in 19, if Lf ′> Lf, that is, if the current number of matching characters is greater than the first longest matching character number, the list including the remaining character strings Si and Sj is replaced with a new character string list SL, The current matching character number Lf ′ is replaced with the first longest matching character number Lf.
Note that in the comparison in step ST19, Lf ′
If <Lf, no processing is performed. In the first character string comparison process from step ST18 to step ST21, all data having the same character from the beginning in character string comparison among all the name data are searched for the name data Di, and the name is searched. All character strings obtained by removing the matching character strings from these name data including the data Di are extracted as node name candidates.

Next, as a second character string comparison process, first, the character string of the name data Di and the character string of the name data Dj are compared from the end, and the number of characters that match continuously from the end is Lb '. The resulting character string is defined as S ′ (step ST2
2). Next, the number of characters Lb ′ is compared with Lb given as the number of characters that have longest matched up to that point in the second character string comparison process for the name data Di (hereinafter, referred to as the second longest matching character number) ( Step ST2
3). In the comparison in step ST23, Lb ′
If> Lb, that is, if the current number of matching characters is greater than the second longest matching character number, the character string S ′ is replaced with the character string S, and the current matching character number Lb ′ is replaced with the second longest matching character number Lb ( Step ST24). In the comparison in step ST23, in the cases other than the above, no processing is performed. Step S above
By the second character string comparison process from T22 to step ST24, all data having the same character from the end in character string comparison among all the name data are searched for the name data Di, and the matching character string is found. It is extracted as a node name candidate.

Next, from step ST13 to step ST
The character string list SL and the character string S obtained as a result of the node name candidate extraction processing for the name data Di executed in 24 are added to the entire node name candidate list NL. In this embodiment, the longest match comparison from the beginning of the character string is executed first, and the longest match comparison from the end of the character string is executed second, but this execution order is reversed. You may.

FIG. 6 is a diagram showing an execution result of the node name candidate extraction processing. The result of executing the node name candidate extraction processing based on the flowchart shown in FIG. 5 for the input data 1 including the plurality of name data in FIG. 2 is shown. 6, reference numeral 21 denotes a step ST shown in FIG.
18 shows the character string extracted as a result of applying the first character string comparison processing of step ST21, that is, the processing relating to the longest match comparison from the head of the character string. For example, for the name data L1 "A condenser vacuum pump inlet valve", the name data with the longest match of characters from the beginning of the character string is L6 "A condenser vacuum pump inlet valve", Since the matching character string is “A condenser vacuum pump inlet”, the extracted character strings are “valve” and “base valve”. In addition, as for the name data L4 “A condenser”, the name data having the longest character match from the head of the character string is L1.
"A condenser vacuum pump inlet valve", L6 "A condenser vacuum pump inlet main valve", L8 "A condenser vacuum pump", and the matching character string is "A condenser" Therefore, the extracted strings are "Vacuum pump inlet valve" and "Vacuum pump inlet main valve".
And a "vacuum pump".

Step ST22 shown in FIG.
22 shows a character string extracted as a result of applying the second character string comparison processing from step 22 to step ST24, that is, the processing relating to the longest match comparison from the end of the character string. For example, as for the name data L1 "A condenser vacuum pump inlet valve", the name data with the longest matching of characters from the end of the character string is L2 "B condenser vacuum pump inlet valve". The character string, that is, the extracted character string, is “condenser vacuum pump inlet valve”.

Reference numeral 23 denotes a node name candidate character string list relating to all node name candidates for the input data 1 obtained by integrating the extracted character strings 21 and 22 to remove duplication. The character string list 23 contains all the names of the class nodes 11 shown in FIG. Therefore, by executing the node name candidate extraction process based on the flowchart shown in FIG. 5, the class node name can be obtained from the name data without the user having to prepare a word dictionary for creating a hierarchical structure in advance. Can be extracted automatically. Note that the character string list 23 includes a character string “Genben” that is not included in the class node illustrated in FIG. 3. There is no problem because it is excluded as a character string to be expressed.

Next, the class hierarchy creation processing in step ST4 shown in FIG. 4 will be described. FIG. 7 is a flowchart showing the processing steps of the class hierarchy creation processing. For each class node name selected by the user in step ST3 shown in FIG. 4, a class node having the class node name is created, and the class hierarchy creation process shown in FIG. 7 is applied to each class node. The class nodes are sequentially added to the hierarchical structure. At this time, the class nodes are processed in ascending order of the number of characters of the class node name.

In this class hierarchy creation processing, the first
, A local variable Np for storing a class node including a root node is prepared, and the root node 14 is stored in Np (step ST31). Next, is_ is added to the node Np.
It is determined whether there is a child node connected by link a (step ST32). If there are child nodes, the subsequent processing is repeatedly executed by the number of child nodes using the counter variable i (step ST33). In this repetition process, the child node is set to Ni, and it is determined whether or not the class node name of the child node Ni is included at the end of the class node name of the current addition target class node (step ST34). . For example, when the class node name of the child node Ni is “entrance valve”, if the class node name of the addition target node is “vacuum pump inlet valve”, “yes”, and the class node name of the addition target node is “vacuum” No for pump outlet valve.

If the determination in step ST34 is yes, the child node Ni is replaced with Np (step ST35), and the process returns to step ST32. On the other hand, if the determination in step ST34 is no, the determination process in step ST34 is executed for the next child node, and if the determination in step ST34 is no for all child nodes, An addition target node is added as a child node of the node Np (step S
T36). Specifically, the addition target node and the node Np are connected by an is_a link. Step ST32
If the determination in step is no, the processing in step ST36 is also performed. In the example shown in FIG. 6, by applying the class hierarchy creation processing shown in FIG. The root node 14 and the class node 11 in the hierarchical structure shown in FIG.
And a partial hierarchical structure consisting of is_a link 13 is created.

Next, the process of adding an instance node in step ST5 shown in FIG. 4 will be described. In step ST5 shown in FIG. 4, an instance node is added to the class hierarchical structure created in step ST4, and h is placed between the instance nodes.
Create an as_a link.

First, with regard to the addition of an instance node, the name data stored in the name data storage means 3 is taken out in the order of the shortest character string, an instance node is created, and the instance node name is changed to the name of the node to be added. The algorithm shown by the flowchart in FIG. 7 is applied again. However, regarding the addition of the addition target node in step ST36 shown in FIG.
Instance_o is defined as the node Np and the node to be added.
It is assumed that they are connected by f link. Then, every time one instance node is added, based on the flowchart shown in FIG. 8, the connection process of the has_a link related to the added instance node is executed.

FIG. 8 is a flowchart showing the processing steps of the has_a link connection processing. In this has_a link connection processing, first, a local variable Nt for storing an instance node is initialized with null (step ST41). Next, using the counter variable i, the subsequent processing is repeatedly executed for all instance nodes (step ST42). In this repetitive processing, first, it is determined whether or not the extracted instance node Ni is the same as the addition target instance node (step ST43). If they are the same, the process moves to step ST42.

If the extracted instance node Ni is not the same as the instance node to be added, it is determined whether or not the instance node name of the instance node Ni is included at the beginning of the instance node name of the instance node to be added ( Step ST44). For example, if the instance name of the instance node Ni is “A
In the case of "condenser", if the instance name of the addition target instance node is "A condenser vacuum pump", it is determined as yes, and if it is "B condenser vacuum pump", it is determined as no. If it is determined as no in step ST44, the process proceeds to step ST42.

If it is determined "yes" in step ST44, it is determined whether or not the number of characters of the instance name of the instance node Ni is longer than the number of characters of the instance name of the instance node stored in the local variable Nt (step ST45). ). In the case of long, Ni
Replace with t (step ST46). If it is short, no processing is performed. As a result, the instance node having the longest number of characters of the instance node name among the instance nodes Ni including the instance name is detected at the head of the instance name of the addition target instance node.

The above steps ST43 to ST
After executing the process of 46 for all the instance nodes, it is determined whether or not the instance nodes are stored in the variable Nt (step ST47). When an instance node is stored in Nt, that is, when Nt is nu
If it is not 11_has_a from the instance node stored in Nt to the instance node to be added
A link is established (step ST48). If no instance node is stored in Nt, no link is established.

In the example shown in FIG.
Are related to step ST5 shown in FIG. 4 with respect to all the name data included in, and the hierarchy creation processing shown in FIG. 7 and has_a shown in FIG.
By applying the link connection processing, all the nodes and links other than the class hierarchical structure created in step ST4 in the hierarchical structure shown in FIG. 3 can be created. In addition, in FIG.
Only the link related to “A condenser” is displayed as the _a link, but the link related to “B condenser” is actually created similarly.

As described above, according to the first embodiment, name data storage means 3 for storing input name data, hierarchical structure storage means 7 for storing a structured database, and name data storage means 3 Node name candidate extracting means for reading the name data stored in the storage device and performing the longest match comparison from the beginning and the longest match comparison from the end with respect to the character strings constituting the name data between the name data to extract the node name candidates 4, node name selecting means 5 for allowing a user to select an appropriate node name, and node name selecting means 5
The node having the node name selected by the above and the node having the name data stored in the name data storage means 3 as the node name are arranged hierarchically, and the mutual relation between the nodes is defined to form the structured database. Since it is configured to include the hierarchical structure creating means 6 for creating, a node name candidate given as a character string representing a candidate of a node to be a component of the hierarchical structure is automatically extracted from the input name data. Since it is possible to automatically create a hierarchical structure using node names and name data selected by the user, it is not necessary for the user to prepare a word database in advance, and to efficiently create a structured database with a small amount of labor. This has the effect that it can be performed.

Embodiment 2 In the structured database creation device according to the first embodiment of the present invention, node name candidates are extracted by comparing character strings between name data. However, due to an input error or the like included in the name data, The correct node name may not be extracted. For example, regarding the name data in the input data 1 shown in FIG. 2, “B condenser” was incorrectly inputted as “A condenser” (that is, two “A condensers” existed). In this case, when the longest match comparison process is performed from the end, the name “A condenser” is extracted and the name “condenser” is not extracted. In such a case, the node name candidate "A condenser" can be discarded using the node name selecting means 5, but the node name "condenser" cannot be created.

FIG. 9 is a diagram showing a configuration of a structured database creating apparatus according to a second embodiment of the present invention which has been made to address such a problem. 9, the same reference numerals as those in FIG. 1 denote the same or corresponding parts, and a description thereof will not be repeated. Reference numeral 31 denotes a hierarchical structure editing unit that enables a user to edit the structured database stored in the hierarchical structure storage unit 7. 32 is a name data storage means 3,
This is a structured database creating apparatus including a node name candidate extracting unit 4, a node name selecting unit 5, a hierarchical structure creating unit 6, a hierarchical structure storing unit 7, and a hierarchical structure editing unit 31.

The hierarchical structure editing means 31 provides a GUI (Gra)
The hierarchical structure stored in the hierarchical structure storage unit 7 is changed in accordance with commands relating to creation, change, deletion, etc. of nodes and links having a hierarchical structure interactively given by the user using a physical user interface). As a result, even in the case described above, the “A condenser” is selected and added to the hierarchical structure and then changed to the “condenser”, or the “A condenser” is discarded and the hierarchical structure is discarded. By creating a new “condenser”, a desired hierarchical structure can be created.

As described above, according to the second embodiment, the structured database stored in the hierarchical structure storage means 7 is provided with the hierarchical structure editing means 31 which enables the user to edit the structured database. Can edit, create, change, delete, etc. nodes and links in the hierarchical structure. Therefore, even if extraction of a correct node name candidate is hindered due to input error of name data, etc., a correct hierarchical structure is created. It has the effect of being able to do so.

Embodiment 3 In large-scale facility data, since a large amount of data is created by a plurality of persons, terms are often not unified. Further, an abbreviated name may be partially used for reasons such as limitation of the number of input characters. For example, "water filling" of "water pump" is described as "water filling", "water filling", "water filling", and "steam generator" is written as "S / G", "SG". Sometimes. When such synonyms are used, in the structured database 8 according to the first embodiment of the present invention, “steam generator”, “S /
G ”and“ SG ”are determined to be separate facilities, and are added as separate nodes to the hierarchical structure.

FIG. 10 is a diagram showing a configuration of a structured database creating apparatus according to a third embodiment of the present invention which has been made to address such a problem. In FIG. 10, reference numeral 41 denotes a synonym definition unit that enables a user to define a synonym using various user interface functions;
Reference numeral 42 denotes a synonym information storage unit that stores synonyms defined using the synonym definition unit 41. Reference numeral 43 denotes the same name data and node name by referring to the synonym information stored in the synonym information storage unit 42. This is a synonym conversion means for converting into a meaningful representative term. Reference numeral 44 denotes name data storage means 3, node name candidate extraction means 4, node name selection means 5, hierarchical structure creation means 6, hierarchical structure storage means 7, synonym definition means 41, synonym information storage means 42, and synonyms. This is a structured database creation device having a conversion unit 43.

The synonym definition means 41 has a function that allows a user to define synonym information interactively in advance or in the process of creating a structured database. FIG. 11 is a diagram illustrating an example of the synonym information. These synonym information is defined using the synonym definition means 41 and is stored in the synonym information storage means 42.

In FIG. 11, reference numeral 45 denotes an identification number of synonym information assigned to each set of synonyms, and reference numeral 46 denotes a list of synonyms constituting the set of synonyms and listing individual synonyms having the same meaning. It shows the definition contents given.
The definition content means that “steam generator” is synonymous with “S / G” and “SG” in the synonym information L101, and “[water-filled] is“ water-filled ”in the synonym information L102. Synonymous with "water filling" and "water filling". As for the underline used in the definition 46, the underlined "steam generator",
"Water-filled" indicates that each is a term representing a set of synonyms.

Next, the operation will be described. The synonym conversion unit 43 refers to the synonym information storage unit 42 for the node name selected using the node name selection unit 5 and the name data read from the name data storage unit 3,
If the node name and the name data or a term partially included in the node name and the name data is registered as a synonym, the corresponding term is uniformly converted to a term representing the synonym.

As described above, according to the third embodiment, the synonym definition means 41 that enables the user to define synonyms, and the synonyms defined by using the synonym definition means 41 are stored. A synonym information storage unit 42; and a synonym conversion unit 43 that converts the name data and the node name into representative terms having the same meaning by referring to the synonym information stored in the synonym information storage unit 42. With such a configuration, even if a plurality of terms having the same meaning are mixed in the input data 1, they can be unified into one term representing these synonyms, and facilities having the same function or specification can be used. And the like can be grouped as the same node on the hierarchical structure, so that an improved structured database can be created.

Embodiment 4 In a large-scale database, various name data are handled, and among them, for example, there is one having the expression “condenser side OO pump”. This means that it is not the accessory equipment of the condenser, but simply the condenser side. If such a name is given, the structured database creating apparatus according to the first embodiment of the present invention will be described. In 8, the longest match comparison from the top with "condenser" extracts a node name candidate "side xx pump". Usually, the user determines that this is not appropriate as a node name (namely, equipment name) and discards it, so that there is no influence on the hierarchical structure creation. However, when many such node name candidates appear as selection candidates, The burden associated with the user's selection work increases. In addition, as a similar expression, there is a phrase such as “for”.

FIG. 12 is a diagram showing a configuration of a structured database creating apparatus according to a fourth embodiment of the present invention which has been made to address such a problem. In FIG. 12, reference numeral 51 denotes an extraction auxiliary rule storage unit for storing an extraction auxiliary rule applied at the time of extracting a node name candidate, and 52 reads the name data stored in the name data storage unit 3 and reads each name data. The node name candidate is extracted by performing the longest match comparison on the character string constituting the name data with all the other name data, and the extraction process of the obtained node name candidate is stored in the extraction auxiliary rule storage means 51. A node name candidate extracting means for discarding the node name candidate when any of the stored rules is satisfied. Also, 53
Is a structured database creating apparatus comprising name data storing means 3, node name selecting means 5, hierarchical structure creating means 6, hierarchical structure storing means 7, extraction auxiliary rule storing means 51 and node name candidate extracting means 52 It is.

The auxiliary rules stored in the extraction auxiliary rule storage means 51 include, for example, [in the longest match comparison from the beginning, when the remaining character strings are “side-” or “use- Do not cut the character string at the part]. Actually, the node name candidate extracting means 52
In the processing of step ST18 shown in FIG. 5, if the extraction auxiliary rule stored in the extraction auxiliary rule storage unit 51 is satisfied, the number of matching characters is set to 0 and the remaining character strings are set to Di, Dj. It is the same. As a result, when the character string matches the extraction assisting rule, even if the character string satisfies the longest matching condition, the character string is rejected at this stage.
In addition, if it is known in advance that the part that is cut off at that part is unnatural as the equipment name, the extraction auxiliary rule storage unit 51 stores an extraction auxiliary rule that prohibits the character string from being cut at that part. It is good to store in.

As described above, according to the fourth embodiment, the extraction auxiliary rule storage means 51 for storing the extraction auxiliary rules applied when extracting the node name candidates and the name data storage means 3 Read name data, perform a longest match comparison on a character string forming name data with all other name data for each name data, extract node name candidates, and obtain the obtained node name candidates. If the extraction process corresponds to any of the rules stored in the extraction auxiliary rule storage means 51, the node name candidate extraction means 52 for discarding the node name candidate is provided. Inappropriate node name candidates can be eliminated before selection by the user, so that the burden of user selection can be reduced and the efficiency of creating a structured database can be increased. There is an effect that it is.

Embodiment 5 FIG. In a large-scale database, as a problem similar to the synonym described in the third embodiment, there is a problem of notation blur. This is the case when the terms in the same katakana notation are written on the one hand in full-width and on the other hand in half-width, or, for example, the term "turbine" is written on the one hand using the long sign "-" and on the other hand "- In this case, a different notation is given as a problem corresponding to the same term.

In such a case, the structured database creating apparatus 8 according to the first embodiment of the present invention performs a half-width katakana and a full-width katakana or a "turbine""
And “turbin” are not determined to be the same node name candidate, but are extracted as separate node name candidates.

To solve such a problem, Embodiment 3
"Turbine" and "turbine" as synonym information described in
Can be dealt with by defining as a synonym. However, according to this method, the long clef "-"
It is necessary to perform the same definition for all terms including, and the burden on the user for synonym definition increases.

FIG. 13 is a diagram showing a configuration of a structured database creating apparatus according to a fifth embodiment of the present invention which has been made to address such a problem. In FIG. 13, reference numeral 61 denotes a name correction rule storage unit that stores a name correction rule that defines a correction method for a term included in name data, and 62 denotes a name correction rule storage unit 61 for each name data included in the input data 1. Is a name correction unit that applies a name correction rule stored in the name data storage unit and accumulates name data converted based on the rule in the name data storage unit 3. Reference numeral 63 denotes a name data storage unit 3, a node name candidate extraction unit 4, a node name selection unit 5, a hierarchical structure creation unit 6, a hierarchical structure storage unit 7, a name correction rule storage unit 61, and a name correction unit 62. This is a structured database creation device to be configured.

As the name correction rules stored in the name correction rule storage means 61, for example, "half-width katakana is changed to full-width katakana" and "long-tone symbols are changed to full-width hyphens" are defined and given. . The name correction unit 62 applies the above-described name correction rule to the name data, for example, changes all half-width katakana to full-width katakana,
In addition, all long sounds are converted to full-width hyphens.

As described above, according to the fifth embodiment, the name correction rule storage means 61 for storing the name correction rule defining the correction method for the term included in the name data.
And applying the name correction rule stored in the name correction rule storage unit 61 to each name data included in the input data 1 and storing the name data converted based on the rule in the name data storage unit 3. Since the configuration is provided with the accumulating name correction means 62, even when the name data has blurring of the notation, it is possible to uniformly correct the notation of the name data by applying the name correction rule. Thus, there is an effect that a more organized structured database can be created. Further, since it is not necessary for the user to perform the work of correcting the notation related to the name data, the burden on the user can be reduced, and the effect of creating the structured database more efficiently can be achieved.

[0065]

As described above, according to the present invention, the name data storage means for storing the input name data, the hierarchical structure storage means for storing the structured database, and the name data storage means Read the name data, and perform the longest match comparison from the beginning and the longest match comparison from the end with respect to the character string forming the name data with all other name data for each name data, and execute the node name. Node name candidate extracting means for extracting candidates, node name selecting means for allowing a user to select a node name of an appropriate node as a constituent element of the structured database from the node name candidates, and a node selected by the node name selecting means When nodes having names and nodes having name data stored in the name data storage means as node names are hierarchically arranged. In addition, since it is configured to include a hierarchical structure creating means for creating a structured database by defining the interrelationship between the nodes, a candidate of a node to be a component of the hierarchical structure is represented from the input name data. A node name candidate given as a character string to be extracted is automatically extracted, and a hierarchical structure can be automatically created using the node name and name data selected by the user. Therefore, the user prepares a word database in advance. There is an effect that a structured database can be efficiently created with no need and with little labor.

According to the present invention, since the structured database stored in the hierarchical structure storage means is provided with the hierarchical structure editing means for enabling the user to edit, the user can create nodes and links for the hierarchical structure, Change,
Since editing such as deletion can be performed, there is an effect that a correct hierarchical structure can be created even when extraction of a correct node name candidate is hindered due to an input error of name data or the like.

According to the present invention, synonym definition means for enabling a user to define a synonym, synonym information storage means for storing synonyms defined using the synonym definition means, and synonym information And synonym conversion means for converting the name data and the node name into representative terms having the same meaning by referring to the synonym information stored in the storage means. Even if a plurality of terms having the same meaning are mixed in the name data, they can be unified into one term representing these synonyms, and facilities having the same function or specification are the same on the hierarchical structure. Since the nodes can be grouped together, there is an effect that a more organized structured database can be created.

According to the present invention, the extraction auxiliary rule storage means for storing the extraction auxiliary rules applied when extracting the node name candidates, and the name data stored in the name data storage means are read, and each name data is read. In each case, the longest match comparison is performed on a character string constituting name data with all other name data to extract node name candidates, and the extraction process of the obtained node name candidates is stored in the extraction auxiliary rule storage means. A node name candidate extracting means for discarding the node name candidate when any of the stored rules is applicable, so that an incorrect node name candidate as a node name is selected before the user selects it. Since the elimination can be eliminated, it is possible to reduce the burden of the user on the selection operation, and to achieve an effect of making the operation of creating the structured database more efficient.

According to the present invention, a name correction rule storing means for storing a name correction rule defining a correction method for a term included in name data, and a name correction rule storing means for each name data constituting input data And a name correction means for accumulating the name data converted based on the name correction rule in the name data storage means by applying the name correction rule stored in the name data. Even if there is blur, the notation of the name data can be uniformly corrected by applying the name correction rule, so that a more structured database can be created. Further, since it is not necessary for the user to perform the work of correcting the notation related to the name data, the burden on the user can be reduced, and the effect of creating the structured database more efficiently can be achieved.

[Brief description of the drawings]

FIG. 1 is a diagram showing a configuration of a structured database creation device according to a first embodiment of the present invention.

FIG. 2 is a diagram illustrating an example of input data including a plurality of name data.

FIG. 3 is a diagram showing a hierarchical structure in a structured database.

FIG. 4 is a flowchart showing a structured database creation method.

FIG. 5 is a flowchart illustrating processing steps of a node name candidate extraction processing.

FIG. 6 is a diagram illustrating an execution result of a node name candidate extraction process.

FIG. 7 is a flowchart illustrating processing steps of a class hierarchy creation processing.

FIG. 8 is a flowchart showing processing steps of a has_a link connection process.

FIG. 9 is a diagram showing a configuration of a structured database creation device according to a second embodiment of the present invention.

FIG. 10 is a diagram showing a configuration of a structured database creation device according to a third embodiment of the present invention.

FIG. 11 is a diagram illustrating an example of synonym information.

FIG. 12 is a diagram showing a configuration of a structured database creation device according to a fourth embodiment of the present invention.

FIG. 13 is a diagram showing a configuration of a structured database creation device according to a fifth embodiment of the present invention.

FIG. 14 is a diagram showing a configuration of a conventional structured database creation device.

[Description of Signs] 1 input data, 2 users, 3 name data storage means, 4 node name candidate extraction means, 5 node name selection means, 6 hierarchical structure creation means, 7 hierarchical structure storage means,
8, 32, 44, 53, 63 Structured database creation device, 11 class node, 12 instance node, 13 is_a link, 14 root node, 15
instance_of link, 16 has_a link, 21 longest match comparison result from the beginning, 22 longest match comparison result from the end, 23 node name candidate character string list, 31 hierarchical structure editing means, 41 synonym definition means, 42 synonym information storage Means, 43 synonym conversion means, 45 synonym information identification number, 46 synonym definition contents,
51 extraction auxiliary rule storage means, 52 node name candidate extraction means, 61 name correction rule storage means, 62 name correction means.

Claims (5)

[Claims] 1. Name data storage means for storing input name data, and a structured database configured by hierarchically arranging nodes corresponding to constituent elements and defining mutual relations between the nodes. A hierarchical structure storing means for storing, and reading the name data stored in the name data storing means, and for each name data, the longest character string from the beginning with respect to a character string forming name data with all other name data Node name candidate extracting means for performing a match comparison and a longest match comparison from the end to extract a node name candidate given as a character string representing a candidate for a node constituting the structured database; The node name candidates extracted by the extraction means are presented, and the node name candidates are used as components constituting the structured database. Node name selecting means for allowing a user to select a node name of a correct node; a node having the node name selected by the node name selecting means; and a node having name data stored in the name data storage means as a node name. And a hierarchical structure creating means for creating a structured database by arranging hierarchically and defining a mutual relationship between the nodes. 2. The structured database creating device according to claim 1, further comprising a hierarchical structure editing unit that allows a user to edit the structured database stored in the hierarchical structure storage unit. 3. A synonym definition means for enabling a user to define a synonym, a synonym information storage means for storing synonyms defined using the synonym definition means, and a synonym information storage means. 2. A structured database creating apparatus according to claim 1, further comprising: synonym conversion means for converting the name data and the node name into representative terms having the same meaning by referring to the synonym information stored in the database. . 4. An extraction assistance rule storage means for storing extraction assistance rules applied at the time of node name candidate extraction, and name data stored in the name data storage means,
For each name data, the longest match comparison from the beginning and the longest match comparison from the end are performed for the character strings constituting the name data with all the other name data, and candidates for nodes constituting the structured database are performed. Node name candidates given as a character string representing the character string. If the extraction process of the obtained node name candidates corresponds to one of the rules stored in the extraction auxiliary rule storage means, the node name candidate is extracted. 2. The structured database creating apparatus according to claim 1, further comprising a node name candidate extracting unit for discarding candidates. 5. A name correction rule storage means for storing a name correction rule defining a correction method for a term included in name data, and a name correction rule storage means for each name data constituting input data. 2. A structured database creation apparatus according to claim 1, further comprising: a name correction unit that applies a name correction rule, and stores name data converted based on the name correction rule in a name data storage unit. apparatus.