JPH11345238A - Method for presenting result of keyword retrieval of html document on www. - Google Patents

Abstract

PROBLEM TO BE SOLVED: To reduce retriever's labor to confirm a retrieval result by deciding a priority with which the retriever should be presented only from cross reference relation in a set of HTML documents of retrieval results. SOLUTION: A retriever 1 accesses a retrieval server computer 4 with a www browser 5 on a client computer 2, inputs a keyword for retrieval and instructs the retrieval. The computer 4 acquires the retrieval keyword through a www server 9 and inputs the retrieval keyword to a keyword retrieval engine 6. The engine 6 produces a list of an HTML document including the given keyword, that is, a 'URL list of retrieval result'. A priority computer 7 calculates a priority about each URL of the produced 'URL list of retrieval results', sorts the 'URL list of retrieval results' in an ascending or descending order based on the result and produces a 'sorted URL list of retrieval results'.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] 1. Field of the Invention [0002] The present invention relates to a technique for searching HTML documents published on the World Wide Web.

[0002]

2. Description of the Related Art A large number of HTML documents are published on the WWW. A full-text search using a keyword is very often used as one of methods for finding necessary information among them. Such a function is called a search engine, and a prominent one is, for example, Alta Vista (http: //altavista.dig) of DEC, USA.
ital.com/).

[0003]

Since a great number of HTML documents exist on the web, tens of thousands to hundreds of thousands of HTML documents containing a certain keyword are often found. Also, it is often difficult to reduce the number of target documents to an appropriate number that can be individually checked even by narrowing down by a plurality of keywords.

[0004] For example, if a search is made using the word as a keyword in order to search for a description of the meaning of a word on the WWW, not only a document explaining the meaning of the word but also all the words using that word are searched. Documents are presented as search results. If you do not know the meaning of the word at all,
It would be very difficult to narrow down the target with yet another keyword, as we cannot expect to come up with another word that might explain the word.

[0005]

Means for Solving the Problems When searching the WWW for an explanation of a word, the result of the conventional keyword search will be related to the possibility that the word is explained as a word. By introducing a scale and reducing the number of result presentations by truncating the scale below a certain value, and by ordering the result presentation based on the values, it is possible to make it easier for the searcher to refer to the results.

One of the features of HTML documents is that HTML documents
The ability to link between documents. The expression form of this link is called a URL. By following the URL embedded in the HTML document using a WWW browser, the linked HTML document can be referred to interactively on the spot.

One of the purposes for which the HTML document was developed is to disclose academic papers. In the presentation of "references", not only the name of the paper / author but also the URL to the HTML document There was the idea of embedding to support access to references.

[0008] As described above, two HTML documents are represented by U
When linked by the RL, it is highly likely that the details of the contents described in the reference source and the explanation of related matters are described in the reference destination.

[0009] Therefore, here, the UR is used as a "measure that is likely to be related to the likelihood of being explained in words".
It is proposed to adopt numerical values related to links between HTML documents by L and to perform ordering and truncation in presentation of search results by the following two numerical values.

(1) For a set of HTML documents obtained as a result of keyword search, U
(2) For a set of HTML documents obtained as a result of keyword search, a set of HTML documents obtained as a result of a keyword search from a reference route of a reference relationship graph by a URL between the documents is referred to. Distance (closer is the higher priority), where (1) is based on the speculation that good explanations will often be cited, and (2) based on the idea of giving priority to the source of the citation.

[0011]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS A detailed embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 shows an example of the overall configuration of a system to which the present invention is applied. The client computer 2 is connected to the search server computer 4 via the network 3. The searcher 1 operates the WWW browser 5 running on the client computer 2 to perform a keyword search of an HTML document on www. In the server computer 4, a keyword search engine 6, a priority calculation mechanism 7, a www page generation mechanism 8, and a www server 9 are operating.

FIG. 2 shows an overall flow of searching for an HTML document including the keyword by a keyword search on the system of FIG. Hereinafter, each step of the overall flow will be described.

Step 1: The searcher 1 accesses the search server computer 4 with the www browser 5 on the client computer 2, inputs a keyword for search, and instructs search. Step 2: The www server computer 4 is connected to the www server 9
Through the keyword search engine 6. Step 3: The keyword search engine 6 creates a list of HTML documents including the given keyword, that is, a “URL list of search results”. Step 12: The priority calculation unit 7 calculates a priority for each URL of the “URL of search result” created in the previous step, and based on the result, “UR of the search result”
The “L list” is sorted in ascending or descending order to create a “URL list of sorted search results”. Step 13: WWW page generation 8 generates a “search result display HTML document” for displaying the “URL list of sorted search results” created in the previous step.
Step 14: The www browser 5 acquires the “search result display HTML document” created in the previous step through the www server 9 and displays it on the client computer 2.

FIG. 3 shows a processing flow of the priority calculation mechanism 7. Parts other than the priority calculation mechanism are conventional techniques, and will not be described in detail here.

Hereinafter, each step of the processing flow of the priority calculation mechanism 7 will be described.

Step 1: "URL list of search results"
Is used to create a “cross-reference map” in steps 1.1 to 1.2. Step 20: Each URL in “URL list of search results”
L for all HTML documents denoted by L.
1.1 to 21 Step 1.1.1: Of all the URLs included in the HTML document, those included in the “URL list of search results” are extracted and set as a “cross-reference destination URL list”. Step 22: The “cross-reference URL list” of all the URLs in the “URL list of search results” is collectively referred to as a “cross-reference map”. -23 Step 2: (priority calculation from "cross-reference map")
For each URL in the “URL list of search results”, “priority calculation” is performed using each “cross-reference URL list”, and a priority value is assigned. Step 24: (Sorting URL list of search results by priority) Sort "URL list of search results" according to the magnitude of the priority value calculated for each URL of "URL list of search results", A “URL list of sorted search results” is created. FIG. 4 is a processing flow of “prioritization by referenced count”, which is an example of the “priority calculation” method performed in the priority calculation mechanism 7. Hereinafter, the processing flow of “prioritization by reference count” will be described.

Step 1: "URL list of search results"
Let K be the number of occurrences of the URL in all the “cross-reference URL lists” constituting the “cross-reference map” for the URL. Step 30: Let K be the priority value of the URL. ~ 3
1 When the method of FIG. 4 is adopted, the one with the larger priority value is prioritized, and the sorting is performed in descending order when sorting by the priority value.

FIG. 5 is a process flow of “prioritization by distance from reference route” which is another example of the “priority calculation” method performed in the priority calculation mechanism 7. Hereinafter, the processing flow of “prioritization by distance from the reference route” will be described.

Step 1: "URL list of search results"
The "distance calculation flag" corresponding to each URL is turned OFF for all URLs. Step 2: All U in "URL list of search results"
For the RL, the priority value is set to + for those that do not have the “cross-reference URL list” (or for which the number of entries is 0).
∞. Step 41: All U in “URL list of search result”
In the RL, “distance from the reference destination route” is calculated for each of those having the “cross-reference URL list” (or the number of entries is one or more), and the value of the result is set as the priority value. FIG. 6 is a processing flow of “calculation of distance from reference destination route” used in processing flow of “prioritization by distance from reference destination route”. Hereinafter, the processing flow of “calculation of distance from reference destination route” will be described.

Step 1: If a "cross-reference URL list" does not exist in the URL, step 2 is executed. If a "cross-reference URL list" exists, step 3 is executed. Step 50: Step 9 is performed with K = 0. ~ 5
1 Step 3: If “URL is being calculated” for the URL
If is ON, execute step 4; if OFF, execute step 5. Step 52: Step 8 is performed with K = + ∞. ~
53 Step 5: Turn on the “distance calculation flag” of the URL
To Step 54: "Calculation of distance from reference route" for each URL on "cross-reference URL list" of the URL
If not completed, the “calculation of distance from reference route” of those URLs is performed. Step 55: Let K = 1 + Max (“calculate distance from reference destination route” value for each URL on “cross-reference list” of the URL). Step 56: OF distance flag is calculated for the URL
Change to F. Step 57: Let K be a calculated value of the distance from the reference route of the URL. 5 and 6, when the methods of FIGS. 5 and 6 are adopted, the one with the smaller priority value is prioritized, and the sorting is performed in ascending order when sorting by the priority value.

[0022]

According to the present invention, in the keyword search of the HTML document, the priority to be presented to the searcher can be determined only from the cross-reference relation in the set of search result HTML documents. Reduce the trouble of checking search results. The validity of the priority based on "reference from other HTML documents" calculated by the present invention is that "links by URL are often used to explain details and related matters". It is not guaranteed to be valid at all times, because it does not exclude its use for other purposes, but it is considered to be much more effective than displaying it at random.

[Brief description of the drawings]

FIG. 1 is a configuration diagram of a keyword search system according to an embodiment of the present invention.

FIG. 2 is an overall flowchart of FIG. 1;

FIG. 3 is a flowchart showing a priority calculation mechanism of the present invention.

FIG. 4 is a flowchart illustrating priority calculation (prioritization based on a referenced count) according to the present invention.

FIG. 5 is a flowchart illustrating priority calculation (priority assignment based on distance from a reference destination route) according to the present invention.

FIG. 6 is a flowchart illustrating calculation of a distance from a reference destination route according to the present invention.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... searcher, 2 ... client computer, 3 ... network, 4 ... search server computer, 5 ... WWW browser, 6 ... keyword search engine, 7 ... priority calculation mechanism, 8 ... www page generation mechanism, 9 ... www server,
10 to 15: Each processing step of the overall flow, 20 to 25
... each processing step of the priority calculation mechanism, 30 to 31 ... each processing step of priority calculation (prioritization based on the referenced count), 40 to 42 ... priority calculation (priority assignment by distance from the reference route) Each processing step,
50 to 59: Each processing step of calculating the distance from the reference route.

Claims (3)

[Claims] 1. A method of searching an HTML document on the web by full-text search for a keyword specified by a searcher and presenting the result to the searcher, wherein a URL in a set of HTML documents obtained by the full-text search for the specified keyword is provided. Characterized by performing ordering based on the state of cross-reference by
A method of presenting a result of a keyword search of an HTML document on www. 2. A method for presenting a result of a keyword search of an HTML document on the web according to claim 1, wherein the ordering method is such that the HTML obtained by full-text search of a designated keyword is used.
H on www, characterized in that the order is based on the value of the referenced count of cross-references by URL in a set of documents
A method of presenting the result of a keyword search of a TML document. 3. A method for presenting a result of a keyword search of an HTML document on www according to claim 1, wherein the ordering method is a method in which the HTML is obtained by a full-text search for a specified keyword.
A method of presenting a result of a keyword search of an HTML document on the www, wherein the order is based on a distance value from a referenced root node in a directed graph of cross-reference by URL in a set of documents.