JP2003271663A - Information retrieval method and device and software - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a technology for retrieving useful information suitable in time among information stored on a network. <P>SOLUTION: When a retrieval keyword is inputted to an information terminal, information retrieval is performed for the server on the network and time information of the information is extracted from the retrieved result for each of the result. The retrieved result is weighted based on the time information, and weight ranking is imparted and the result is outputted. <P>COPYRIGHT: (C)2003,JPO

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information search method and device on a network. In particular, it relates to a technique for searching effective information at the time of searching.

[0002]

2. Description of the Related Art In recent years, commercialization of the Internet and Intranet has progressed, and the amount of information stored therein has increased dramatically. A service called a search site is provided to search information from this enormous amount of information, and a search engine for high-speed and effective search is under development.

For example, a WW called Google (trademark)
According to the W service, "Google is designed to make information scattered in an orderly manner. It is not a compilation or limited directory, or a list of expensive search results sold, but according to the unique structure of the Internet. "We are organizing the information on the Web and realizing the ideal form of a search service."
It discloses that a technology called ageRank (trademark) is used.

That is, "PageRank (TM) takes advantage of the characteristics of the huge link structure of the Web. The link from page A to page B is regarded as a support vote for page B by page A, and Google Determines how important the page is by this number of votes, but Google simply says
So not only do you see the number of links, but you also analyze the pages that voted. The votes cast by the "important" pages are more highly evaluated, and the pages that receive them are "important". High PageRa for important pages that have been highly evaluated by such analysis
nk (TM) (page rank) is given, and the rank in the search results is also high. PageRank (TM) is a comprehensive index that shows the importance of a page on Google, and is not affected by each search. Rather, PageRank (TM) is a property of each web page itself, based on an analysis of the link structure according to complex algorithms. ", The method of determining the importance of the search result centering on the number of links is used.

The above-mentioned conventional method is extremely effective in weighting the search results, that is, in weighting, and information that is actually linked in large numbers is automatically or artificially linked by a person who has already referred to the information. Since it is often judged to be useful, it is likely to include the necessary information. However, according to only this method, there is a problem that sufficient search cannot be performed for information such as tourist information whose validity is limited in time.

That is, for example, when a tourist information is searched for a special opening of a Buddhist statue in a temple, many links are made by the conventional method if it is the noted opening, and the opening period ends. However, since the link is not always deleted, it will continue to be output as a search result of high importance. However, in this example, if the disclosure period is apparently ended, the information is useless for a searcher who is looking for tourist information, which rather deteriorates the search efficiency.

Although search results may be preferentially output from newly recorded ones in the past, they do not judge the time suitability of the information itself, and particularly suit the time required by the searcher. It was not a technique for weighting the search results for those that were done.

[0008]

SUMMARY OF THE INVENTION Therefore, the present invention is intended to solve the problems of the above-mentioned prior art, and it is useful information that is temporally adapted from the information recorded on the network. The purpose is to provide a search technology.

[0009]

In order to achieve the above object, the information retrieval method of the present invention has the following features.
That is, when a single or a plurality of information storage servers are provided on a network such as the Internet or an intranet, it is a method for searching information from an information terminal connected to the network, which includes at least the following steps. (1) A step of generating a search keyword in an information terminal, an information search is performed on at least one of an information storage server based on the search keyword, or a search information storage server that has previously collected search information from the information storage server Step (2) extracting time information of the information for each result from the search result of the information search, (3) relating to at least one of the search keyword and the search execution time based on the time information And weighting the search results, and (4) ranking and outputting the search results based on the weighting.

In the present invention, unlike the prior art, the search results are weighted based on the relationship between the time information of the search results and the time desired by the searcher to search for useful information that matches in time. You can

The time information may be a time at which the search result is recorded, updated, or deleted in at least one of the information storage server and the search information storage server.

The time information specifies a character string indicating the time information in advance, and when the character string matches the character string included in the search result by a predetermined threshold value or more, extraction is performed. It may be time to be.

Further, according to the present invention, it is possible to provide an information retrieving apparatus for retrieving information from an information terminal connected to the network when one or more information storage servers are provided on the Internet or intranet network. The apparatus includes at least one of a keyword generation unit that generates a search keyword in an information terminal, an information storage server based on the search keyword, or a search information storage server that collects search information in advance from the information storage server. An information search unit that performs information search, a time information extraction unit that extracts time information of the information for each result from the search results of the information search, and based on the time information, at least one of the search keyword and the search execution time. And a search result output unit that weights the search result in relation to the search result and a search result output unit that ranks and outputs the search result based on the weight.

The time information may be a time at which the search result is recorded, updated, or deleted in at least one of the information storage server and the search information storage server, or a character string indicating the time information is designated in advance. The time may be the time for checking for a match with the character string included in the search result and automatically extracting it.

Here, in the configuration in which the information retrieval method / apparatus according to the present invention retrieves a process when performing one or a plurality of processes within a designated period, the time information is a time required for the process itself. First time information, second time information that is a time required to perform the process other than the process, and third time information that is another time required in the designated period,
It can be classified into period information indicating the designated period.

Then, when any of the time information is obtained from the first generated search keyword, a search keyword for obtaining at least one of the other time information is automatically generated to obtain the other time information. At the same time, the search results may be weighted in relation to at least one of the search keyword and the search execution time based on a plurality of time information.

In particular, in the structure in which the period information is the travel period, the first time information is the event time at the destination, the second time information is the access time to the place where the event is performed, and the third time information. The time information may be a necessary / occurring living time in daily life.

[0018]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the examples shown in the drawings. The embodiment is
Modifications can be made without departing from the spirit of the present invention. FIG. 1 shows a block diagram of an information retrieval device (1) according to the present invention. The device (1) uses the Internet (2)
Search for desired information from an infinite number of information servers. Note that the present apparatus (1) can be used in the completely same configuration not only on the Internet but also on an intranet configured by only a limited range of servers and terminals.

The present apparatus (1) has, as main components, a CPU (4) for performing arithmetic / communication processing, a clock unit (6) for giving time such as current time to the CPU, and an external hard disk for storing information. A recording means (7) is provided. Then, the keyword generation unit (10) of the CPU (4) generates a keyword to be searched from the information input from the input device (3) such as a keyboard.

Regarding the generation of the keyword, a word or phrase directly input from the input device (3) may be used, another word or phrase may be automatically added, or a related word may be extracted by a thesaurus database provided separately, or a language translation may be performed. Alternatively, the character type may be converted to enable more effective search.

For example, in the case of a device for retrieving sightseeing information, even when the searcher inputs "Nara" with the input device (3), words such as "sightseeing" and "travel" are added. Generate keywords. Further, "Nara" may be converted into "nara" or "nara". As described above, according to the present invention, the keyword is generated in the keyword generating unit (10),
Not only the input by the keyboard but also free keywords can be used based on the information given from other devices.

Information is collected from the information recording server (20) which is connected to the Internet (2) and records information, or the Internet (2) in advance by using a keyword generated by the keyword generating unit (10), for recording and recording. The information retrieval unit (11) retrieves matching information from the retrieval information storage server (21). The search information storage server (21) may not be connected via the Internet (2) as shown, but may be directly connected to the device (1) as indicated by a dotted line part (22). Alternatively, both paths may be provided.

Various known search techniques can be used for the search in the information search section (11). In the present invention, the information to be searched is included in the record date, update date, and delete information of the information. In the system where is recorded, it is desirable to acquire the date of deletion as a search result. Further, in the present invention, since time information is important, it is necessary to search for new information as much as possible, and the search information recording server (2
When the search speed is increased by 1) or the like, it is desirable that the update frequency is high.

The search results are temporarily stored in the search result database (30) provided in the external recording device (7). At this stage, they are stored in the same rank as the search results by the known search technique, and are sorted by, for example, the matching rate of the keywords or the rank according to the importance according to the number of links shown in the above-mentioned conventional technique.

According to the present invention, in order to output time-adapted information from these conventional techniques, the following time information extraction unit (12) and search result evaluation unit (13) are newly added to center the time axis. The search results are weighted. First, the time information extraction unit (12) will be described.

The time information extraction unit (12) checks the search results in the search result database (30) whether or not time-related information is included, and if so, extracts the information. In the method of extracting, for example, the recording date or the update date of the information in the information recording server (20), for example, the html file in the WWW (World Wide Web)
Focus on date information. These pieces of information are important indicators of the freshness of the information, and the usefulness of the information is judged from the absolute passage of time, rather than simply arranging the new ones as seen in the conventional method. Also enables.

In a system such as a search information recording server (21) that obtains information in advance and uses it for a search, since the amount of information on the Internet has increased, the cache information and the actual time when the system accessed the search target site There may be differences in the information on the sites that exist in. Therefore,
In addition to the above, the system search time may be extracted as time information.

As the time to be extracted by the time information extraction unit (12), the information description time explicitly described such as "update date" included in the search result can be used. Conventionally, for example, date information in an html file often indicates the date on which the information was artificially created or modified, but recent WWW sites use an access counter or the like to monitor access to the site. Programs often work automatically, and the date information is often rewritten accordingly. Therefore, it can be considered that the information explicitly described in the information is more reliable than the information on the system such as the date information is extracted.

The character string indicating the recorded / updated time is prepared in advance in the time designation character string database (31) of the external recording device (7), and the matching character string is searched from the search result. This configuration is artificially entered rather than the information on the system, and can be extremely effective information for judging the freshness of the information.

Further, it is possible to search the information which is considered to indicate the time information from the search result. That is, for example, in order to extract the date and time of an event, the time designation character string database (31) is provided beforehand with character strings such as "date", "period", and "open", and these are searched. When included in the result, the time information designated by the instruction is extracted.

Here, the above-mentioned time information is not limited to the time information of the event such as "date", "period", "opening", and may include other time concepts. In summary, when performing one or more processes within a certain period (period information), the time required for the process itself (first time information),
Time (second time information) required for preparation / movement, etc. in order to perform the processing, and other time (third time information) that is approximately restricted within the period such as meal time, sleep / rest time, or sunshine time Time information such as is considered.

When applied to a trip, the event holding time is the first time information, and the access time required for transportation to the event is the second time information, the traveler's wake-up time, meal time, break time, etc. Living time can be classified as third time information.

Further, even in the case of searching for an event which is the purpose of information search, the present invention allows searching for the access time and the living time. That is, for example, when the first time information is extracted by the time information extraction unit (12), the keyword generation unit (10) generates a keyword for extracting the access time to the event place.
For example, when the museum opening time is extracted, the keyword generating unit (10) generates a keyword "museum train route" for searching a route to the museum, and the information searching unit (11) searches.

As a result, the time required for the train, etc.
You can get time information. Similarly, if it is necessary to be bright when visiting the park, the third time information can be extracted by searching for a keyword such as "park sunset".

Based on the time information extracted by the time information extraction unit (12) as described above, the search result database (3
The search result in (0) is weighted by the search result evaluation unit (13). At this time, the device (1) includes the clock unit (6), and the searcher obtains the search time at which the search is performed, so that the search result can be weighted in comparison with the time information. Further, the keyword generation unit (10) may include the optimum time in advance in the keywords, and the search results may be weighted by comparison with these keywords.

For example, the following technique can be considered as an example of using the date of recording or the date of updating in the information recording server (20). That is, for a year-old information, a predetermined light weight may be uniformly applied, or a weight may be set for each elapsed time, and such information is recorded in the external recording device (7). Can be set.

Further, it is an event that is further developed and held regularly. For example, information related to "Hatsumode" is set to the lightest weight for all information recorded / updated more than a year ago. Alternatively, the weight to be set may be changed according to the word / phrase included in the search result. By equipping the device (1) with a database in which a large amount of information relating to the temporal validity of these pieces of information is provided, it is possible to achieve more practical output of search results.

When the information description time explicitly described such as "update date" is used, weighting is performed depending on whether the update date is suitable for the search time or the keyword. For example, the highest weight is set when the update date is within one week from the search time, and conversely, the weighting is performed stepwise so that the information that is 6 months or more has the lowest weight.
As a result, the freshness of highly reliable information such as "update date" can be strongly reflected in the search result.

In such a structure, a search result having no character string indicating the recorded / updated time in the search result may be weighted lower than a search result having a character string. If there is no explicit description, there is no means to measure the freshness of the information, and it is unsuitable for obtaining a time-matched search result as in the present invention. Therefore,
Lower the weight or use the information on the system, such as da
Weighting may be performed in combination with te information or the like.

Further, in the case of searching for time information from the search result, it is often possible to determine a more specific time match such as the event holding period, and the weighting effect is particularly high. If these events and the like do not match in time and the output as a search result is small, set an extremely small weight, and conversely, if they match the search time or keyword, set the maximum weight. It is possible.

When investigating the presence or absence of temporal matching, the concept of time described above can be combined and evaluated.
For example, even if only the event time is matched, it is often not matched when combined with the access time, and the first time information or the third time information is combined with the period information to judge whether or not the searched keyword matches. Good. As described above, the search results can be ranked in consideration of the temporal matching that is a feature of the present invention.

In the search result evaluation section (13), each search result is weighted, and the search result database (3
Rank the search results in 0). According to the order, the search result output unit (14) outputs a suitable search result from the device (1). The output search result may be displayed on an output device (5) such as a monitor, or may be used on another processing device.

Although the present invention has the above-mentioned configuration, the present invention is not limited to the above-described embodiment, but is also provided as a search method that can be used as an algorithm of search software. As the search device, software may be incorporated in a personal computer to have the same configuration as described above, or the search device may be provided as a dedicated terminal integrally provided with an input device and the like.

The information retrieval target is not limited to the above WWW html, and any information can be used. For example, a configuration may be used in which a database that stores information distribution services using electronic mail is referenced, and the date of reception and the included character string are referenced.

[0045]

Since the information retrieval method and apparatus of the present invention have the above-mentioned configuration, they have the following effects. That is,
It is possible to retrieve useful information that is temporally suitable from the vast amount of information recorded on the network, which can contribute to improvement in retrieval efficiency. In particular, since it costs access cost to collect information through the network, according to the present invention, it is possible to suppress cost increase due to repeated access to unnecessary information and repeated search.

[Brief description of drawings]

FIG. 1 is a block diagram of an information search device according to the present invention.

[Explanation of symbols]

1 Information retrieval apparatus according to the present invention 2 Internet 3 input devices 4 CPU 5 Output device 6 clock section 7 External storage device 10 keyword generation section 11 Information Retrieval Department 12-hour information extractor 13 Search Results Evaluation Department 14 Search result output section 20 Information recording server 21 Information Recording Server for Search 22 Connection route 30 Search result database 31 time specified character string database

   ─────────────────────────────────────────────────── ─── Continued front page    (72) Inventor Yasuyuki Kono             10-6-401 Oide Town, Nishinomiya City, Hyogo Prefecture (72) Inventor Masatsugu Kido             Nara City Nara City School Asahimoto Town 2-1903-33 F term (reference) 5B075 KK03 NK02 NK31 NR15 PP22                       PQ02 PQ75 PR08

Claims (10)

[Claims] 1. A method of performing information retrieval from an information terminal connected to the network when a single or a plurality of information storage servers are provided on the Internet or intranet network, wherein a search keyword is generated in the information terminal. A step of performing an information search on at least one of an information storage server based on the search keyword or a search information storage server that has collected search information from the information storage server in advance; Extracting time information of the information for each result, weighting the search result in relation to at least one of the search keyword and the search execution time based on the time information, the search result based on the weighting At least each step of the ranked and output step Information retrieval method which is characterized in that there was example. 2. The information search method according to claim 1, wherein the time information is a time at which the search result is recorded, updated, or deleted in at least one of the information storage server and the search information storage server. 3. The time information specifies a character string indicating time information in advance, and when the character string matches a character string included in the search result by a predetermined threshold value or more, extraction is performed. The information retrieval method according to claim 1 or 2, which is a time to be performed. 4. The configuration wherein the information retrieval method is a method for retrieving one or more processes within a designated period, wherein the time information is a time required for the process itself. The time information, the second time information that is the time required to perform the process other than the process, the third time information that is the time required for the other process within the designated period, and the period information that indicates the designated period. When one of the time information is obtained by classifying the search keyword generated first, the search keyword for obtaining at least one of the other time information is automatically generated to obtain the other time information, 4. The information search method according to claim 1, wherein the search result is weighted in relation to at least one of the search keyword and the search execution time based on a plurality of time information. 5. In a structure in which the period information is a travel period, the first time information is an event time at a destination, and the second time information is an access time to a place where the event is performed, The information search method according to claim 4, wherein the third time information is a life time required / occurred in daily life. 6. An information retrieval device for performing information retrieval from an information terminal connected to the network when having a single or a plurality of information storage servers on the network of the Internet or an intranet, wherein the retrieval keyword is used in the information terminal. A keyword generation unit that generates a keyword, an information search unit that performs information search on at least one of an information storage server based on the search keyword, or a search information storage server that has collected search information from the information storage server in advance; A time information extraction unit that extracts time information of the information for each result from the search results, and weights the search results in relation to at least one of the search keyword and the search execution time based on the time information Search result evaluation unit, a search result that ranks and outputs search results based on the weighting. Information retrieval apparatus characterized by comprising at least the result output section. 7. The information search device according to claim 6, wherein the time information is a time at which the search result is recorded, updated, or deleted in at least one of the information storage server and the search information storage server. 8. The time information is a time when a character string designating the time information is designated in advance and a match with the character string included in the search result is checked and automatically extracted. 7. The information retrieval device described in 7. 9. The configuration in which the information search device searches for a process when performing one or more processes within a designated period, wherein the time information is first time information that is a time required for the process itself. , Second time information, which is the time required to perform the process other than the process, third time information, which is the other time within the designated period, and period information, which indicates the designated period, When any of the time information is obtained from the first generated search keyword, the keyword generation unit automatically generates a search keyword for obtaining at least one of the other time information, and the information search unit and the time information extraction Other time information is obtained by the section, and at the same time, in the search result evaluation section, the search result is related to at least one of the search keyword and the search execution time based on the plurality of time information. 9. The information search device according to claim 6, wherein weighting is performed. 10. The structure in which the period information is a travel period, the first time information is an event time at a destination, and the second time information is an access time to a place where the event is performed, The information search device according to claim 9, wherein the third time information is a life time required / occurred in daily life.