KR100433584B1 - Method for product detailed information extraction of internet shopping mall with ontology and wrapper data - Google Patents

Abstract

The present invention relates to a method for extracting detailed information about an internet shopping mall product using ontology and rule information.

According to the present invention, a preprocessing step of converting an HTML document obtained by using a document collecting robot into a document composed of purely non-tag strings and <TR, <P, and <BR tags based on shopping mall product detail page URL information; An ontology-based retrieval step of extracting a classification name and an ontology value of a product based on ontology information of a knowledge database in the document converted through the preprocessing step; By extracting only the model name of the product from the document converted through the preprocessing step, detailed information as well as general simple information of the product may be extracted.

This makes it possible to obtain detailed information about the product accurately and quickly, which simplifies work that requires a lot of manpower and time.

Description

Method for product detailed information extraction of internet shopping mall with ontology and wrapper data}

The present invention relates to a method of extracting detailed information on an internet shopping mall product using ontologies and rule information. In more detail, the present invention relates to HTML based on the ontology information of a knowledge database and the rules of a wrapper. The present invention relates to a method for extracting detailed information about an internet shopping mall product using ontologies and rule information for quickly and accurately extracting detailed information about a product on a document.

If a user wants to get some information from a plain text document or a document obtained through a user query on the Internet, most users read the document and learn how to get the information they want. think. If you have to read all the documents on the web and grasp the information yourself, you will have to spend a lot of time. However, if the document is in the form of a database, or if it is a semi-structured document that is not in the form of a complete database, the user can automatically extract the necessary information without reading all the documents. This technique of obtaining information from a large number of documents available on the Internet web or documents stored as strings but not on the Internet web is called information extraction.

In such information extraction, if there is a rule common to the document to be obtained information, this rule can be used to obtain information easily. In general, most documents on the Internet web are constructed using HTML tags. In other words, if HTML documents have a certain form and information is expressed by using various kinds of tags, it is possible to easily extract necessary information by finding the rules in the document. For example, if a document always contains the document's title between the <B> and <B /> tags, the rule automatically extracts the title from the document. Also, even if it's not made up of HTML tags, You can also automatically extract information from a document using rules such as the word title followed by the title. In order to extract the information automatically through this method, it is necessary to be able to recognize the rules existing in the document. This is called a wrapper. In other words, by using a wrapper, necessary information can be automatically extracted from a document conforming to the rules of the wrapper. The information extraction method using only such a wrapper is the most frequently used method in the existing information extraction method. However, if the format of the wrapper is determined by the type of the document, there are no rules in the document, and the provided document changes from time to time, the composition of the wrapper for the document is not only possible, but the information extracted from the document is not available. The disadvantage is that it is impossible to extract based on semantics.

Another method commonly used in the information extraction method is to use a dictionary such as ontology. This method compares and searches all data in the dictionary such as ontology in the document to obtain information, and extracts any data that exists in the dictionary. However, this method not only requires a large dictionary for information extraction, but also consumes a lot of time for searching the system.

In addition, the biggest disadvantage of the existing methods for extracting the product information of the Internet shopping mall using the wrapper or the ontology is that the form of the extracted information is simple. This is because the HTML documents of the Internet shopping mall have different forms, so it is difficult to use a wrapper, and even if a wrapper is used, only a simple wrapper can be used. In addition, most systems using Ontology have a drawback in that it takes much time to execute by inputting only a representative URL of a shopping mall and collecting all related HTML documents and analyzing them.

An object of the present invention for solving the problems of the prior art described above is to quickly and easily extract detailed information of a defined product using ontologies and rule information, thereby quickly and easily extracting various product information of an internet shopping mall. An object of the present invention is to provide a method for extracting detailed information about an internet shopping mall product using topology and rule information.

1 is an overall flowchart of a method for extracting detailed information about an internet shopping mall product using ontologies and rule information according to an embodiment of the present invention;

2 is a flowchart illustrating a preprocessing process of converting an input HTML document into a form for extracting information;

3 is a flowchart illustrating a method of extracting ontology information based on data of a knowledge database of goods in a preprocessed document;

4 is a flowchart illustrating a method of extracting only a model name from detailed information of a product.

The present invention for achieving the above object is a method for extracting detailed information about an internet shopping mall product using ontologies and rule information through a detailed information extraction system having a knowledge database and a database. A first step of finding an HTML page related to the internet shopping mall product using the corresponding URL information; A second step of preprocessing the HTML page into a document consisting of specific tags and pure strings; A third step of extracting detailed information of the shopping mall product except for a model name from the preprocessed document using the ontology and synonyms in the knowledge database and storing the detailed information of the shopping mall product in the database: the model name based on rule information of the preprocessed document And extracting only the bay and storing it in the database.

Preferably, a record storing a program executable by a computer to perform a method of extracting detailed information about an Internet shopping mall product using ontologies and rule information through a detailed information extraction system having a knowledge database and a database. A medium, comprising: a first step of finding an HTML page related to the Internet shopping mall product by using corresponding URL information of the shopping mall product; A second step of preprocessing the HTML page into a document consisting of specific tags and pure strings; A third step of extracting detailed information of the shopping mall product except for a model name from the preprocessed document by using the ontology and synonyms in the knowledge database and storing it in the database; And a fourth step of extracting only the model name on the basis of the rule information of the preprocessed document and storing the model name in the database.

Hereinafter, a method of extracting detailed information about an internet shopping mall product using ontologies and rule information according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings. 1 is an overall flowchart of a method for extracting detailed information about an internet shopping mall product using ontologies and rule information according to an embodiment of the present invention.

First, a URL (Uniform Resource Locator: Internet address or domain name, hereinafter referred to as URL) information of a corresponding shopping mall for which the detailed information extraction system administrator wants to extract detailed information is received (S110). Using the received URL information of the shopping mall, the system manager extracts the product information HTML document from the shopping mall by using a document collecting robot (S120), and retrieves the retrieved HTML document from the ontology and model name. The document is preprocessed as a search (S130). The preprocessed document first searches for detailed information using the classification name and ontology of the product entered in the knowledge database (S140), while searching for a model name for extracting a particularly basic model name from the detailed information of the product. Perform the step (S150).

At this time, the purpose of the step (S130) of pre-processing the imported HTML document as a document for ontology search and model name search, to remove all tags except <TR, <P, <BR, and pure In order to reduce the amount of data to be searched by reducing the amount of data to be retrieved by configuring a document consisting of a string and a tag irrelevant to the HTML tag.

In addition, after separately specifying the step of extracting only the model name in the present invention, using only the wrapper (WRAPPER), the model name is the most basic unit of the detailed information of the product, which can use the ontology value like other detailed information. This is because the rules exist on the product HTML page in the same manner as the method of the present invention.

2 to 4 are flowcharts showing the detailed operation of the method for extracting detailed information about an Internet shopping mall product using ontologies and rule information according to the present invention. With reference to the drawings it will be described in detail with respect to the method for extracting detailed information about the Internet shopping mall products using the ontology and the rule information according to the present invention. 2 is a flowchart illustrating a preprocessing process of converting an input HTML document into a form for extracting information.

First, after separating the HTML document found through the document collecting robot in the detailed information extraction system line by line (S210), it is searched whether each line in the HTML document separated by line includes the HTML tag character (S220). If the search results include HTML tag characters in each line, the corresponding tag line including HTML tag characters is detected (S230), while if the search result does not include HTML tag characters, the tag does not include HTML tag characters. After searching the lines (S270), the lines including the searched non-tag characters are stored in the corresponding database (S280). On the other hand, if the corresponding line including the HTML tag character is detected, it is searched whether the corresponding line includes specific tag characters of <TR, <P, <BR (S240). As a result of the search, if the line does not include the specific tag character as described above, the corresponding line is removed (S250), while if the search result is the line including the specific tag character as described above, the corresponding line is stored in the database (S260). ). In this manner, after storing the corresponding lines in the database, it is searched whether a line to be input and converted remains (S290). As a result of the search, if the line to be input still remains, the above process is repeated, otherwise, the preprocessing step is terminated.

3 is a flowchart illustrating a method of extracting ontology information based on data of a product knowledge database from a preprocessed document.

First, the detailed information extraction system finds the classification names related to the extraction target product and the synonyms of the classification name from the knowledge database of the product (S301), and then reads the corresponding HTML document created as a result of preprocessing line by line (S302). The synonyms closest to the purpose are searched by comparing the synonyms and the strings of the corresponding classification names in the read line (S303). As a result of the search, the detailed information extraction system searches whether the corresponding synonym to find exists in the read line (S304). As a result of the search, if the synonym exists, the ontology values related to the classification name are loaded from the knowledge database (S305), and then the ontologies are searched by comparing the loaded ontology values with the front and back of the string (S307). On the other hand, if a corresponding synonym does not exist as a result of the search, all ontology of the search product is searched (S307). Next, it is determined whether an ontology value exists (S308). As a result of the search, if the ontology value exists, the classification name and the found ontology value are stored in the database (S309). If the search result, the ontology value does not exist, all ontology for the search product is stored. Search (S307). As such, when the classification name and the found ontology value are stored in the database, it is searched whether a line to be input and searched still remains (S310). If the search result, the line to be entered and search still remains, the above process is repeated, while if the search result, there is no line to be entered and searched, the ontology search step ends.

4 is a flowchart illustrating a method of extracting only a model name from detailed information of a product. After reading the preprocessed document line by line through the flow chart shown in FIG. 2 (S410), it is searched whether there is a character "(" in the read line (S420). If a character exists, the character string between "(" character and ")" is extracted (S430), and then the character string is searched for the existence of the "-" character from the extracted string (S440). (If there is no character, it is searched whether a "-" character exists in a corresponding line read immediately (S440). If a "-" character is present, the search based on the "-" character (S430). After separating the extracted string from (S450), it is analyzed whether the separated character is composed of a combination of alphanumeric characters (S460), and as a result of the analysis, if the separated character is composed of a combination of alphanumeric characters, the separated character Is stored in the database as the model name (S470), and still Also, it is determined whether the character strings to be processed remain (S480), and as a result of the determination, the character strings to be processed remain and the above steps are repeated, while if the character strings to be processed are not found, the model name of the search product is searched. On the other hand, if there is no "-" character in the read line, or if the separated character does not consist of a combination of alphanumeric characters, it is searched whether there are still strings to be processed (S480). ).

While the invention has been described above based on the preferred embodiments thereof, these embodiments are intended to illustrate rather than limit the invention. It will be apparent to those skilled in the art that various changes, modifications, or adjustments to the above embodiments can be made without departing from the spirit of the invention. Therefore, the protection scope of the present invention will be limited only by the appended claims, and should be construed as including all such changes, modifications or adjustments.

As described above, according to the present invention, by using the ontology-based data and the synonym with the extended meaning, it is possible not only to accurately extract the detailed information of the search product, but also to search for the target HTML document based on the target URL information. By shortening and deleting unnecessary tag information by a preprocessing process, the amount of data to be processed can be reduced, thereby reducing the time for detailed information extraction.

Claims (6)

A method for extracting detailed information about an internet shopping mall product using a detailed information extraction system having a knowledge database (a) and a database (b), A first step of searching for a HyperText Markup Language (HTML) page related to the Internet shopping mall product using corresponding URL (Uniform Resource Locator) information of the shopping mall product; A second step of preprocessing the retrieved HTML page into a document composed of specific tags and pure strings; A third step of searching for synonyms close to the synonyms of the classification name and a string in the input line of the preprocessed document from the classification names and synonyms stored in the knowledge database (a); A fourth step of determining whether a corresponding ontology value exists by comparing the ontology values stored in the knowledge database (a) with the corresponding character string if the synonyms found in the third step exist in the read line; A fifth step of storing and updating the classification name and the ontology value in the database (b) if a corresponding ontology value exists as a result of the determination in the fourth step; And A sixth step of extracting detailed information on the Internet shopping mall by using the stored and updated database (b); Method for extracting detailed information about an Internet shopping mall product, comprising a. The method of claim 1, The second step, Sub-segmenting the HTML page by line and then removing the HTML tag except for a specific tag by the separated line; And Storing and updating a line including only the specific tag and a line not including the HTML tag in the database (b); Method for extracting detailed information about an Internet shopping mall product, comprising a. delete The method of claim 1, A second step of searching for a "(" character and a ")" character for each line by reading the preprocessed document in the second step; Extracting a string between the searched "(" and ")" characters, and then searching for a "-" character in the extracted string; And After separating the character strings based on the searched "-" character, if the separated character strings are composed of a combination of alphabets and numbers, recognizing the character strings as model names, and storing and updating the character strings in the database (b). -3 step; Method for extracting detailed information about an Internet shopping mall product, comprising a. The method of claim 1, And the specific tag includes <TR, <P, and <BR tags. A recording medium having recorded thereon a program capable of executing a computer-implemented method of extracting detailed information about an Internet shopping mall product using a detailed information extraction system having a knowledge database (a) and a database (b), A first step of searching for a HyperText Markup Language (HTML) page related to the Internet shopping mall product using corresponding URL (Uniform Resource Locator) information of the shopping mall product; A second step of preprocessing the retrieved HTML page into a document composed of specific tags and pure strings; A third step of searching for synonyms close to the synonyms of the classification name and a string in the input line of the preprocessed document from the classification names and synonyms stored in the knowledge database (a); A fourth step of determining whether a corresponding ontology value exists by comparing the ontology values stored in the knowledge database (a) with the corresponding character string if the synonyms found in the third step exist in the read line; A fifth step of storing and updating the classification name and the ontology value in the database (b) if a corresponding ontology value exists as a result of the determination in the fourth step; And A sixth step of extracting detailed information on the Internet shopping mall by using the stored and updated database (b); A computer-readable recording medium having recorded a program comprising a.