JPH02224069A - Information retrieving system - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed Description of the Invention] [Industrial Application Field] The present invention relates to an information retrieval system, and in particular, for example, in order to simultaneously output text information of a search result and image information that is semantically related, a plurality of pieces of information can be output. This invention relates to an information retrieval system that performs simultaneous searches.

[Conventional technology]

There have been various types of information retrieval systems that include image information, etc., but conventional systems of this type include, for example,
As described in Information Processing Society of Japan, June 1987, p. 699, this is a format in which an inquiry is made by specifying search conditions. In other words, this is a search that focuses on image information with specific content.

In addition, while there are many techniques for searching text information that use index structures, in recent years, associative techniques using hardware have become popular, as described in the Journal of the Information Processing Society of Japan, June 1986, page 593. Memory research is also becoming more active.

Furthermore, regarding a document display system for a document consisting of a text page and a figure page, Japanese Patent Application Laid-Open No. 1518-1983
The following technology disclosed in Japanese Patent No. 90 is known. That is, when there is a reference description of a figure using a figure number in a page of text I-, the correspondence between the page of that text and the page of the referenced figure is stored in advance, and the text I-
When the page is displayed, the figure is also displayed at the same time.

[Problem to be solved by the invention]

However, the above-mentioned conventional technology does not take into consideration the point of displaying (outputting) other auxiliary information at the same time to help understand the information content of the search results.
However, the above-mentioned third prior art is only capable of displaying figures and reference pages at the same time if they are associated with each other in advance.

The present invention has been made in view of the above circumstances, and its purpose is to simultaneously display other auxiliary information closely related to the basic information that is the search result. An object of the present invention is to provide an information retrieval system that facilitates understanding of search results.

[Means to solve the problem]

A double object of the present invention is to provide an information retrieval system that outputs first information as a search result together with second information related to the content of the first information. a first knowledge base that stores the information, a second knowledge base that stores the semantic content of the second information, a knowledge base for connecting between the first and second knowledge bases, and a knowledge base that stores the search results. This is achieved by an information retrieval system characterized by having means for searching each knowledge base and outputting the first information together with second information related to the meaning and content of the first information when the first information is obtained. .

[Operation] In the information retrieval system according to the present invention, for example,
When the first information is text information and the second information is image information, there is a semantic relationship between the knowledge base that stores the semantic content of the text information and the knowledge base that stores the semantic content of the image information. Since the speed test is performed using a knowledge base such as a thesaurus and synonyms to search for image information related to text information, the desired image information can be reliably found.

[Example] Hereinafter, an example of the present invention will be described in detail based on the drawings.

FIG. 1 is a block diagram showing the overall configuration of an information retrieval system that is a second embodiment of the present invention.

In the figure, reference numeral 101 indicates a text information file, which stores information that is the subject of an inquiry input from the timing input mechanism 103. Further, 102 is a knowledge base that stores the semantic content of the information in the text information file 101, 104 is an image information file, and 105 is knowledge that stores the semantic content of the information in the image information file 1.04. 6106 showing the base
From the semantic content of the information in the text information file 101 and the semantic content of the information in the image information file 104, the requested text 1 =, the syntax used to search for the information.
・Knowledge base regarding words such as ras and synonyms 1, 107 is a search mechanism for searching the text information file 101, 1
08 is a text display mechanism that displays information in the text information file 101 as a search result; 109 is a search mechanism that searches for an image information file 104 related to the information in the text information file 1-01; and 110 is a search mechanism for the image information file 104. Display mechanism 111- and Jl-2 have text information file 1018 and image information file 1
This is a mechanism that displays the information of 04 side by side.

Note that the search mechanism 107. The text display mechanism 108, search mechanism 109, and image display mechanism 110 are generally realized by computer programs.

FIG. 2 shows examples of the various knowledge bases described above. 201 shows an example of the text information file i-01. Here, the content (semantic content) of the text information "User purchases product Y to improve productivity" is stored as knowledge in the knowledge base 202 as shown in the figure. It shows. The structure of the knowledge base 202 is an expression method called a case structure, and is written mainly using verbs. The indicates the upper-lower relationship. 204 is an example of an image information file, and 205 is a knowledge base that stores the meaning and content of the image information as knowledge.

FIG. 3 shows a flowchart 1 of the search process in the first embodiment configured as described above. Below, Figure 1~
The operation of the first embodiment will be explained based on FIG.

Here, given the text information, image information, and knowledge base shown above, we will calculate the productivity rate using product Y.
An example of a case where the query "I want to see the second result" will be explained below.

When the above-mentioned inquiry is input to the above-mentioned make-order input mechanism 103, the inquiry is accepted in step 301, and information such as ``ryl 1 item on make-order'' and ``improved productivity'' is received.

Focusing on "to aim", the timing is 6 20283 in Figure 2
? 2024 F Transformed into a shape equivalent to 2025 7
Nine. In step 302, the knowledge base 2Q'2 is searched using the modified alignment. Since this item overlaps with the 20.23, 2024.2025 (7) portions, the search conditions match, and the text information file 2°1 corresponding to this knowledge is obtained.

In step 303, this text information is displayed on the text display screen 11 via the display mechanism 108 described above. In step 304, the image information knowledge base 205 is searched using the portions 2021-2025 in the knowledge base of the text information. This search is
Start with "Purchase" of 2051-, then r of next 2052
Replace company 1 with thesaurus "user", 2055
By replacing "product" with the specific [Y product].

Each part of 2051.2052.2055 is 20
Since it overlaps with the part of 21.2022.2025, it is considered that there is a related part.

In step 305, image information corresponding to the found knowledge, ie, 204, is extracted. In step 306, the extracted image information is displayed on the image display screen 112 via the display mechanism 110 described above. In Figure 4,
The display results are shown together with the timing input contents. In the figure, 40
Reference numeral 1 indicates the content of the manual adjustment, 402 indicates an example of text display, and 403 indicates an example of image display.

FIG. 5 shows the display drawing 5 by scrolling the display screen 402.
01, and a new image information screen 502 corresponding to this text information is displayed.

In addition, FIG. 6 shows table information 60 regarding the round table when the "Round Table" part of the text information is specified on both sides of FIG.
2 is displayed.

The underline of the round table in the text screen 601 indicates this. This is an indication of an intention to order that the display content of the image information be different from the usual display.

In the first embodiment above, an example was shown in which the first information is text information and the second information is image information. A second embodiment in which the information is sentence structure information of the text information will be described.

In the second embodiment described below, an example will be shown in which a list of concise sentences including headwords is displayed as the second information.

FIG. 7 is an example of text information to be searched in the second embodiment. In an information retrieval system in which a large amount of such text information is stored, for example, consider a case where text information including the headword "Knowledge Base Management System" is searched. There is usually a large number of text information that satisfies such search conditions, and displaying the entire text for each piece of text information as shown in 701 in Figure 7 would result in too large a display amount of search results. . Therefore, it becomes necessary to display text information partially. The purpose of this embodiment is to display search results effectively in such cases by treating words that have a case relationship with the headword as important words and displaying them mainly around the headword. FIG. 8 schematically represents structured information that explicitly expresses the modification relations and case relations of the words in one sentence in the above-mentioned text i/information 70. In FIG. 8, 801 is word information corresponding to a clause in the above-mentioned text information 701, and 802 is structural information representing case relationships and modification relationships between words. The lower end connected by a line indicates the modifier, and the upper end indicates the modified word.

The case relationship between words indicates that the element connected by the arrow 802 is a case-dominant word, and the former is a case-dominant word.

FIG. 9 is a block diagram showing the overall configuration of the information retrieval system of the second embodiment, and corresponds to FIG. 1 shown above. Terminal device 902 in this second embodiment
The input mechanism 103 and the display mechanism 111 in FIG.
112, 901 is a storage device that stores structured information consisting of the above-mentioned word information and structure information, and 903 stores structured information stored in this storage device 901 and a terminal device. 902 shows a display information creation device that creates information to be displayed as a search result based on search conditions input from 902.

FIG. 10 is a diagram showing the storage format of structured information stored in the storage device 901. In the figure, 1001 is a structured information storage table.

This structured information storage table]001 consists of records 1002 corresponding to each clause in the text information. Each node F L OO2 is the record identifier 10.
03, 1004 independent words in a bunsetsu, 100 attached words in a bunsetsu
5. A modified flag 1006 indicating the presence or absence of a clause that modifies the clause represented by the relevant record, a modified optical identifier 1008 representing the identifier of the record corresponding to the clause modified by the clause represented by the relevant 1-no code, and their case relationships 1. OO7
It consists of fields that represent

The content of the modified flag 1006 is 111n when the clause is modified by another clause, and is "0" when it is not modified. Case relation field 1007 stores It O11 if there is no modification target, and if there is a modification target, it is stored from the case dominant side to the case dominant side in the direction of the modification between bunsetsu clauses and the I blood decoration. If you are heading towards the opposite side, use a plus sign (+).
, if the direction is opposite, a negative sign (-) is added and a value representing the path relationship is stored.

Moreover, a special delimiter, for example, a record in which "-1" is stored in the modified flag field 1006, is inserted between each piece of text information.

When the storage device 901 receives a search command, it compares the headword specified in the search command with the value of the independent word field 1o04 of each record 1-002.

If there is a matching record, include that record,
The l/code group within the range bounded by the record representing the break between text information is transferred to the display information creation device 903.

FIG. 11 is a diagram showing a work area used by the display information creation device 903. ]-101 indicates a structured information editing area in which a group of records transferred from the storage device ff 901 mentioned above is stored.

This structured information editing area 1101 is stored in the storage device 901.
It consists of record 1]02, which is made up of finodes similar to the structured information storage table.

Further, 1.103 is a display information storage area in which created display information is stored. The display information storage area 1103 is
Field 110 for storing headwords that are search conditions
5. It consists of a field 1104 for storing a character string to be displayed before the headword and a field 1106 for storing a character string to be displayed after the headword.

The operation of this embodiment configured as described above will be explained below with reference to FIGS. 12 and 13. In addition, Fig. 12 shows
FIG. 13, a flowchart for explaining the operation of the above-mentioned display information creation means 903, is a diagram showing the display on the above-mentioned terminal device 902.

In FIG. 13, 1301A and 1301B indicate display states when searching is performed using different search conditions. In Fig. 13, reference numeral 1302 designates the entry word to be searched for, and 1303 indicates a peripheral display level input that specifies the peripheral display level that indicates how far the area around the entry word is to be displayed following the route. Canilia, 1304 This is a display-excluded road-to-road entry canilia that specifies a road-to-road worker to be excluded from display when displaying the vicinity of the entry word.

1.301A represents a search condition for displaying text information that includes "knowledge base management system" as a headword, excluding modal cases and locative cases, and tracing the road attendants around the headword up to four levels. There is.

The operation will be explained in detail below.

In step 1201, the search conditions shown in 1.301A are read from the terminal device 902. In this example, "knowledge base management system" is read as the headword, "modal case" and "locative case" are read as the 41 display exclusion relations as the peripheral display level, and the display information storage area 1103 of the headword is read.
It is stored in field 1105 of . Subsequently, in step 1202, a search command is issued to the aforementioned storage device 901 using the headword read in the previous step 1201. In step 1203, as described above, a group of records containing text information including records including headings and words are transferred from the storage device 901, and these records are stored in the structured information editing area 1101. .

In the next step 1204, it is determined whether or not the record group was transferred in step 1203. While the record group has been transferred, the search results are displayed in the following steps 1205 to 1210, and the record group is transferred. and when all the displays are completed, the entire process ends.

After storing the one-note group corresponding to the text information to be displayed in the structured information editing area 11o1, step 12
In step 05, the record containing the headword in the structured information editing area 1101 is located and set as the target record. Also, fields 1104 and 1106 are cleared.

In the next step 1206, starting from the target record 1102 located in the previous step 1205, the following processing is recursively executed for the target record. First, if the value of the modified flag 1006 is dg 1 $1, that value is set to ``''O+t, and the record + 724 identifier 1003
Records containing the modified optical identifier 1008 are traced. Hereinafter, this trace destination record will be referred to as the qualification source record. The road attendant 1007 of the modification destination record determines whether or not it is the display excluded case relation input in step 120], and if it is the display excluded case relation inputted in step 120], if it is the display excluded case relation, the modification source 1 code 1
The modified optical identifier 1008 of 102 is updated to II OII, and if there is no display exclusion case relationship, the modified flag ]00G of the modified record is set to 111.11 again. Next, assuming that the target located in the previous step 1205 is the O level, the number of levels traced through the records to the modification source record is (C)-L, and the number of labels must reach the peripheral display level input in step 1201. For example, the modified flag 1006 of the modification source record is updated to 'O', and this step ends.

If the peripheral display does not reach 1 novel, the modification source record will be newly targeted! This step 1206 is executed recursively as a code.

In step 1207, the target record road attendant 1007 starts from the target record located in step 1205, traces the record in the opposite direction to step 12σS, and updates the modified optical identifier j-008. It is determined whether or not it is the input display exclusion case relation, and if it is the 1 display exclusion case relation, the modification light identifier 1008 is updated to u On. At this point, the modified optical identifier 100
8 is "On," or the record located in step 1205 is set as 0 level, and the number of levels traced to the target record is counted, and the number of levels reaches the peripheral display level input in step] 201. If so, proceed to the next step 1208. At that point, if the modified optical identifier 1008 is not "0" and the number of levels traced through the record has not reached the above-mentioned peripheral display level, the target record The record having the value of the modified optical identifier 1008 as the record identifier 1003 is traced, and this record is set as the target record in this step 12.
Execute 07 again. Note that the contents of the structured information editing area 1101- shown in FIG.
Based on the contents of o1, create a character string biography to be displayed before the headword. Therefore, in step 1208, the record containing the headword located in step 1205 is set as the target record, and the following process is recursively executed for the target record.

If the modified flag 1o06 of the target record is 111 Tj, the record having the value of the identifier 1003 of the target record as the modified optical identifier 1008 is traced.

This tracing source record will be called a qualification source record. If there are multiple records to follow.

The independent word 1004 and attached word 1005 of the modification source record are subjected to primary processing in descending order of the record identifier, and are filled into the field 11o4 from the rear. and,
Step 1208 is executed recursively using the modification source record as a new target record.

Next, in step 209, a character string to be displayed after the headword is created based on the contents of the structured information editing area 1101. Therefore, in this step 1209, the record containing the entry word set as location 1 in step 1205 is set as the target record, and the following processing is executed for the target record.

If the modified optical identifier 1008 of the target record is Li OIT, this step is ended and the process proceeds to step 121.0. If you drop it with '0'', the value will be set to record identifier 100.
3, the independent word 1004 and attached word 1005 of the traced record are concatenated, and the field 1106 is filled from the front, and this step 1209 is recursively executed using the traced source record as a new target record.

Finally, in step 1210, 1 display information storage area 1
The display information in 103 is transmitted to the terminal device 902, and the display information in 1.
Figure 3 (7) 130], A, 1301 B
Display as shown in d. In the search result display areas 1305 of 1301A and 1301B, the 100th position of two headwords is displayed with an asterisk (*). In the example of 1301Δ, the character string before the headword is 42 characters, I F O),
, In the example of H, 18 characters are 21-original script information 7
01, the 42 characters before the headword are ``we have adopted an inference engine that derives the underlying information from the explicit data stored in the search engine'', and the 18 characters before the headword are ``information to do''. We adopted an inference engine that guides
6 In other words, according to this embodiment, by selectively displaying important parts centering on headwords, it is possible to display search results that are easy to understand and effective. 9 In the embodiment, if the search conditions input in step 1201 are stored in the storage device 901, and the designation of the headword, peripheral display level, or display excluded case relationship is omitted in subsequent search condition inputs, It is also possible for the display information creation device 903 to read them from the storage device 901 and use them instead of the input from the terminal device 902.

Furthermore, in the second embodiment, the structured information as shown in FIG. 10 created by the user on the display screen of the terminal device 902 can be stored in the storage device 90. be. This makes it possible to store text information as structured information.

In the above embodiment, text information is used as the information to be searched, and image information or sentence structure information of the text information is exemplified as the information displayed additionally. It goes without saying that the present invention is not limited to these, but can be used to search and output various other types of information.

The third embodiment will be described below.

FIG. 14 shows a flowchart of processing in the third embodiment, in which first information, second information whose semantic content is related to this first information, and whose semantic content is related to this second information. It describes the operation of repeatedly searching and retrieving related information like the third information. FIG. 14 corresponds to FIG. 3. Figures 14 to 17 below
The operation of the third embodiment will be explained with reference to FIGS. In FIG. 14, in step 1401,
Accept impromptu search conditions. The search conditions are shown in Figure 4 40
The content is shown in ]. In step 1402, information matching the search conditions is retrieved and displayed. FIG. 4 402 is an example. This is the first information. Step 14
In step 03, second information related in meaning and content to the first information is searched and extracted using the knowledge base of the first information and the knowledge base of the second information. In step 1404, the retrieved second information is displayed. FIG. 4 403 shows an example. In step 1405, the user determines whether or not to retrieve third information related to the second information. If ejecting is to be performed, step 140
Execute 6. In this step 1406, third information related in meaning and content to the extracted second information is searched and extracted using the knowledge base. Figure 16 160
1 is a slightly scrolled view of 403 in FIG. 4, and the words "Used in the S factory to be constructed" are visible.

In Figure 4 403, the display that appeared on the right side of the screen is the 16th
In FIG. 1601, it is displayed on the left side of the screen, but this is because the position of 403 in FIG. (right side) to the position of 1601 in Fig. 16 (
(left side). Fig. 16 1601. The knowledge base for this is shown in FIG. 15 500. This shows more parts than FIG. 2 205.

1502 to 1507 in FIG. 15 are larger than those in FIG. 2 205. The text information knowledge base 1508 is searched using the portions 1502 to 1507 in FIG. As a result of the search, the parts 1501, 1502, and 1503 in FIG.
Since it can be seen that it overlaps with the part 12, the part 1513 in FIG. 15 is extracted as the related part. With the above processing, step 1406 ends and the process moves to step 1404. In step 1404, the extracted third information (151 in FIG.
3) is displayed. In step 1405, the user further determines whether or not to retrieve the fourth information related to the third information.If step 2 is to be performed, the same process is repeated.If not, step 14
The process moves to 07 and ends. FIG. 17 shows FIGS. 4 and 5.

This figure shows the relationship between the second information related to the first information shown in FIGS. 6 and 16, and the third information related to the second information. 1701 is 402 in FIG. 4, 501 in FIG. 5
．． Figure 6 601. represents the first information shown in . 170
3 tree 402.1704 *501, ]-705 zero 6
01 corresponds to 401, 501, and 601, respectively. Although 501 and 601 are the same text information, the related second information is different, so they will be described separately here. Generally, there can be more than one piece of first information.

Reference numeral 1702 indicates this, but since it is the same for the specific examples shown in FIGS. 4 to 6, it is omitted. 1706, 1710. L713 represents the second information, and tree 403 of 1708, *502 of tree 1601.1712 of 1709, and *602 of 1715 correspond to 5, respectively, 403.1601, 502, and 602. The arrows 1720.1721.1722 indicate that the content of the tree 402 in the first information 1701 is the second information 1,70.
6 (D related to 403 d, *50 in 1701)
1 is related to *502 in 1710, and zero 60 in 1701 is related to *602 in 1713. 1707 indicates that there is one more piece of information related to the book 402.

Similarly, 1711 indicates another related information for *501, and 1714 indicates another related information for *601. 1716 represents the third information, and *1602 represents 1602 in FIG. 1
The arrow 723 indicates the second information *1601 in 706
1717 indicates that the content of the third information 1716 is related to the content of the book 1602. Steps 1404, 1405 in FIG. By repeatedly performing step 1406, as shown in FIG. 17, second information related to the first information and third information related to the second information are extracted one after another. Flowchart 1 shown in FIG. 14 shows an operating principle that can be executed when there is one piece of first information, one piece of second information, one piece of third information, etc. As shown in Figure 17, if there is a plurality of first information, second information, third information, etc., or if the number of items increases further due to scrolling, a more complicated flowchart than that shown in Figure 14 is required. It is. FIG. 18 is a flowchart for this purpose. The explanation will be made while comparing FIG. 18 with FIG. 14. Step 1801 is the same as step 14o1 for accepting an inquiry. Step 1802 is the same as step 402 in that information matching the search conditions is retrieved. However, when there is a large number of pieces of information, it is necessary to memorize the information and the management data for managing it.
Display only the first information on the screen table. This corresponds to 402 in FIG. 4 and 1703 in FIG. 17. Management data will be described later. Step 1803 is step 140
Same as 3.

However, management data is also stored at the same time. Steps 1404 to 1.407 in FIG. 14 are step 1 in FIG.
The details are shown in 804-1816. Refinement is a detailed description of the controls for when there are multiple pieces of information that match the search conditions, when there are multiple pieces of information that have related semantic content, or when scrolling to extract information that has related semantic content. . Step 1804 is a step of displaying the first item of information extracted in step 1803 on the screen table. Figure 4 403, Figure 17 1708
corresponds to Step 1805 prompts the user to decide whether to scroll the screen table.

If Yes, in step 1806, scrolling is performed, information related in meaning and content to the information displayed on the screen table is extracted, and the information and management data of the information are stored. Then, the first information is displayed on the screen table. Fifth
This corresponds to FIGS. 501, 502 and 1704 and 1712 in FIG. In step 1807, related information is displayed by specifying a specific location 601, 602 in FIG. , FIG. 17 1705
, 1-715. In step 1808, when there is plurality No. 1 information, the user is asked whether or not to display the next information displayed on the screen table. If Yes, the following information is displayed. This corresponds to 1702 in FIG. Thereafter, the process of step 1806 is performed. In step J-810, the user is asked whether or not to further display information whose semantic content is related to the information on the screen table that is displayed as information whose semantic content is related. If Yes,
In step 1811, the display information of the screen table is transferred to the screen table,
The process of step 1806 is performed. This corresponds to 1601 and 1602 in FIG. In step 1812, the user is asked whether to scroll the information on the screen table and display the next part. If Yes, the information is scrolled and displayed in step 1813. In step]-814, when there are multiple pieces of information on the screen table, the user is asked whether or not to display the next piece of information. If yes, step 8
15 displays the next information. In FIG. 17, when 1706 is displayed on the screen table, this corresponds to displaying 1707 on the screen table. Management data will be explained with reference to FIGS. 19 and 20. 2001 to 20o9 in Figure 20 are 1703, 1704, 1705, and 1 in Figures 1 to 7, respectively.
702,1708,1.7o9゜1712.1715,
This is management data for 1718 information. The management data consists of six data items shown in 1901 to 1906 in FIG.

1601 is a number indicating -, 2, etc. of first information, second information, etc. 1602 is a line number indicating the position in the information. 2011 indicates that it is the 10th line. This indicates that the sentence starting with "user is" in FIG. 4 402 starts from the 10th line. 1903 is, for example, the first
This shows that the next position after 1703 in FIG. 7 is 1704.

In 2012, the address q i 701 where the management data 2002 is stored is the data item value. 1
904 indicates an address of information for management data.

2013 is the address a17 where 17o1 is stored
1905, which represents 01, is the address to the next information when multiple pieces of information exist. 2014 is the management data 2゜O, 4゛(i&, address is p1702) of the following information.
It shows that. 1906 is an address to information related to semantic content. 2015, the address of the management data 2005 of information related to semantic content is p17
I remember that it was 03.

The fourth embodiment will be described below.

FIG. 27 is a flowchart 1 of processing according to the fourth embodiment.
This is a flowchart for returning to the previously viewed information or proceeding from the information viewed on the left side of the screen and the right side of the screen using the management data explained in Figures 19 and 20. It is. Step 2101 corresponds to steps 1801 to J804 in FIG. 18.

This step 2]01 is a step in which the address of the management data corresponding to the information displayed on the right side of one screen of the five-screen table is stored. Step 2103 corresponds to steps 1805 to 1814 in FIG.

Steps 2104 to 21 are steps for O to go back or forward. Step 2104 is
The user is asked whether or not to move forward or backward regarding the scroll wheel or specific point estimation for the currently displayed information. If Yes, in step 2105, the control data is used to move forward or backward; in FIG. 20, When the management data 20o2 is the management data of the currently displayed information, when backward scrolling is instructed, the backward address p1701 of 2o]6 is looked at, and the management data 200] is looked up.

Then, information for 2001 is displayed. 2106 is
If YES in the forward/reverse timing for multiple cases, step 21o7 is executed. In Figure 17, 1702
to 1701.

This corresponds to a movement such as from 1706 to 1707.

Step 2108 is a timing for returning from the second information to the first information or proceeding to the second information, and if Yes, step 2109 is executed. 1st
, 7, from 1717 to 1,706, from 1706 to 1
This corresponds to a movement such as to 701 or from 1701 to 1706. When 1.717 is displayed on the screen table,
When one step forward is specified, 1706 is displayed on the screen table.

C. Effects of the Invention] As described above, according to the present invention, in an information retrieval system in which the first information of the search results is output together with the second information related to the content of the first information. , a first knowledge base that stores the semantic content of the first information, a second knowledge base that stores the semantic content of the second information, and a link that exists between the first and second knowledge bases. When the knowledge base and the first information of the search results are obtained, the system is configured to have means for searching each knowledge base and outputting the first information together with the second information related to the meaning and content of the knowledge base. , it is remarkable that it is possible to realize an information retrieval system that helps the understanding of search results by simultaneously displaying other auxiliary information closely related to the basic information that is the search result. It can be effective.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the overall configuration of an information retrieval system according to the first embodiment of the present invention, FIG. 2 is a diagram showing an example of the knowledge base of the first embodiment, and FIG. A flowchart of the search process in the first embodiment, FIGS. 4 to 6 are diagrams showing display examples of the first embodiment, and FIG. 7 is a flowchart of the search process in the second embodiment. Illustration showing an example, No. 8
The figure is a diagram schematically representing the structured information of text information of the second embodiment, FIG. 9 is a block diagram showing the overall configuration of the information retrieval system of the second embodiment, and FIG. FIG. 11 is a diagram showing the storage format of structured information in the second embodiment; FIG.
2 is a flowchart 1. for explaining the operation of the second embodiment, FIG. 1r3 is a diagram showing an example of display on the terminal device of the second embodiment, and 14. [1'; i is the search flow of the third embodiment - char - 1...], Figure 5 shows an example of the knowledge base of the third embodiment, and Figure 16 shows the example of the knowledge base of the third embodiment. A diagram showing a display example of the third embodiment, FIG. 1.7 is a diagram showing the relationship of display information amount in the third embodiment, and FIG. 18 is a flowchart showing screen operations in the third embodiment. , FIG. 19 and FIG. 20 are diagrams showing system management data for storing the history of screen operations in the third embodiment, and FIG. 21 is a diagram showing the previous screen when double-sided operation is performed in the fourth embodiment. This is a processing flowchart that makes it possible to perform the following steps. 101... Text information, 102... Knowledge base regarding the semantic content of text information, 103... Intermediate input mechanism, 1o4... Image information, 105... Knowledge base regarding the semantic content of image information, 106 ...
Knowledge base regarding words, 107...Search mechanism, 10
8...Text display mechanism, 109...Search mechanism, 4
02...Text display example, 403...Image display example, 901...Storage device, 902...Terminal device, 9
03...Display information creation means, 1.301A, 1301
B...Structured information display example, 1401 to 1406...
Flowchart for repeatedly searching related information, 2101~
2111: Flowchart for going back and viewing searched information. ro diagram flute diagram zo diagram De DI Akume// v7/z diagram hand/fu diagram Nepa diagram

Claims (1)

[Claims] 1. In an information retrieval system that outputs first information as a search result together with second information related to the content of the first information, the semantic content of the first information is A stored first knowledge base, a second knowledge base that stores the semantic content of the second information, a linking knowledge base interposed between the first and second knowledge bases, and a search result. an information retrieval system comprising means for searching each of the knowledge bases and outputting the first information together with second information related to the meaning and content thereof when the first information is obtained. . 2. The information according to claim 1, further comprising means for outputting, when the first information of the search results is scrolled, second information regarding the meaning and content of the first information after scrolling. Search system. 3. The method further comprises means for outputting the second information related thereto by specifying an arbitrary location in the first information of the search result. The information retrieval system described in 2. 4. According to claim 1, further comprising means for displaying, in the first information, which part of the first information the outputted second information is related to. information retrieval system. 5.Furthermore, when there is a plurality of pieces of the second information related to the first information in the search results, the display position of the related part in the first information is sequentially forwarded according to an instruction. 5. The information retrieval system according to claim 4, further comprising means for changing the information retrieval system. 6. Claim 1, wherein the first information is text information and the second information is image information.
2. The information retrieval system according to any one of 4 and 5. 7. The first information is text information, and the second information is sentence structure information of the text information. Information retrieval system. 8. The information according to any one of claims 1, 2, 4 and 5, wherein the linking knowledge base interposed between the first and second knowledge bases is a knowledge base regarding words. Search system. 9. Further, when the second information related to the semantic content of the first information of the search result is obtained, an operation of obtaining third information related to the semantic content of the second information; 2. The information retrieval system according to claim 1, further comprising means for repeatedly obtaining related information such as an operation for obtaining fourth information related to the semantic content of the third information. 10. Furthermore, information that is repeatedly output such as outputting the second information related to the semantic content of the first information and outputting third information related to the semantic content of the second information,
2. The information retrieval system according to claim 1, further comprising means for returning to and outputting previous information.