KR20000036758A - A method for establishing database for searching files and a method for searching file by use of the database - Google Patents

Abstract

PURPOSE: A method for searching texts with use of a search database of texts is provided, that a searcher using a computer predicts texts to be used in searching, extracts the texts, and divides the extracted texts into parts according to a meaning, to extract a predicted searched text as a sentences. The method is provided, that the searcher establishes a search database by making a the texts as a database by use of the extracted text. The method is provided, that the searcher divides sentences expressed a search propose directly, into parts according to the meaning, and searches the texts through the established database by inputting the texts as the sentences. The method is provided, that the searcher searches the inputted texts, to search texts of the search database by use of the searched predicted texts. CONSTITUTION: A method for searching texts by use of a search database of texts contains processes. The first process is to input basic information to texts. The second process is to extract a predicted text from an inputted corresponded text, to divide the extracted text into parts according to a meaning, and to input the predicted text as a sentence, which is composed of many parts. The third process is to convert the inputted data into an inner data type. The fourth process is to input the converted data in a search database, as a record type proper each field of the database.

Description

A method for establishing database for searching files and a method for searching file by use of the database}

The present invention provides a document data search database and a document data search method using the same, wherein a person searching for document data using a computer predicts a search sentence that can be used at the time of the search in advance and extracts a search sentence into a plurality of parts according to meaning Extracts the expected search sentence in the form of sentences composed of parts, builds a search database by documenting the document data using the extracted predicted search sentence, and the searcher has a sentence that can directly indicate the intention to search. According to the search results in the form of a sentence composed of a plurality of parts divided into parts to search the document data through the built search database, and the searcher searched the expected search text entered in the search database Document in the search database using One to allow the search to be fast and efficient search, while easy-to-use document data retrieval relates to databasing and document data retrieval method using the same.

The most widely used data search in the current computer use may be Internet data search. Internet data search is to find a website that contains necessary document data among websites scattered on the Internet. Currently, search engines use search engines on how to manipulate and manage information about websites in search programs. Currently commercialized search engines search for many websites on the Internet and include desired keywords (keywords). It is used by indexing the web site's address (URL) into its main word (keyword) and storing it in a database to serve those who use the Internet. In other words, when a conventional Internet user wants to search for data on the Internet, the user inputs a URL (Uniform Resource Location) of a search engine site that provides an Internet information search engine on the Internet, and then searches for the search screen of the connected search engine site. By entering a search expression for information and accessing the information related to the search expression among the data scattered on the Internet by using the relationship with the search expression entered, the user provided the information suitable for the search expression to the Internet users. . Conventionally used Internet data search engines include a directory method, a robot method, and a meta method. The directory method is a method in which a user or an operator enters Internet information by hand and divides it into directories according to their own classification method. When a directory method is registered, the address of a website and a description of representative contents of the website are displayed. This is because the entire document contents of the website are not registered, but the representative address of the website and its representative contents are registered. The robot method is a robot, not a human, that registers a website automatically as it travels the Internet. Therefore, no input of a user or operator is necessary and is automatically updated at regular intervals. Unlike the directory method, the contents of all documents are stored, resulting in tens of millions of data. The meta method does not have its own data or search engine, and if a person who wants to search enters a search expression, the search expression is sent to other search engines and searched to receive the results from the search engines, sorted, and regularly sorted. This is how it is shown.

In the above-mentioned internet data search engine, the directory method is searchable by field, and the searched data is likely to match the information to be searched for, and it searches a website about a topic, not a document containing specific information. Is useful, but has the disadvantage of not being able to search in detail. In addition, the robot method searches web sites by document, which makes it possible to search for a specific word. However, the robot method has a disadvantage in that an excessive amount of unnecessary data that is far from the content to be searched is searched for. There is a disadvantage that it is difficult for beginners to use because the search requires a sufficient prior knowledge of the search expression. In addition, the meta-method has the effect of using multiple search engines at the same time, and the search results are also excellent, but it takes a long time, so it is currently rarely used. In addition, in the conventional Internet data search engine, the searcher cannot directly use the expression indicating the purpose of the search as a search sentence, and the search should be performed by creating a search term or a search expression on the contents that can represent the contents of the website to be searched. This can cause inconvenience in the process of performing a search, and it is not easy to find a search term that is representative of the contents of a website to be searched if the searcher does not know the contents of the searched material in detail. In particular, it is a disadvantage that the searching process is not very easy, especially for beginners who are not used to internet searching. In addition, when searching the Internet materials, the search is performed by creating contents that are representative of the contents of the website to be searched in the form of search terms or search expressions. Although it is different from the search results, there is a disadvantage that it is necessary to re-search among the search results or the search results are not used properly because the searched data is too large.

In the same way, the Internet data search engine can be used to search document data shared through an Intranet that builds an internal information system of a company and a specific organization using the Web technology of the Internet. Therefore, there is a disadvantage in that it has the same problem as described above.

The present invention has been made to solve the above problems, a sentence that is composed of a plurality of parts divided into parts according to the meaning by extracting the search sentence in anticipation of a search sentence which can be used by a person searching for document data using a computer in advance. Extract the expected search sentence in the form, construct the search database by documenting the document data using the extracted predicted search sentence, and the searcher divides the sentence that can express the intention to search directly into parts according to the meaning. Enter the search sentence in sentence form consisting of two parts to search the document data through the built-up search database. Also, the searcher searches the expected search text entered in the search database and uses the searched expected search text. Allows you to search documentation in Document data retrieval database for easy, fast and efficient retrieval and document data retrieval method using the same.

1 is a diagram showing a flow of data of the present invention.

2 is a flow chart illustrating a document data search database process.

3 is a flowchart illustrating a process of a searcher searching for document data through a search database.

4 is a flowchart illustrating a process of a searcher searching for an expected search sentence through a search database.

Figure 5 is a flow chart of an embodiment showing the process of the similarity rate calculation step.

FIG. 6 is a diagram illustrating an embodiment of a screen for inputting basic document information and an expected search statement to construct a search database. FIG.

FIG. 7 is a diagram illustrating an embodiment of a screen displaying input results so that basic information of a document and an expected search statement are inputted into a search database.

8 is a diagram illustrating an embodiment of a screen provided with a search sentence input unit for a searcher to input a search sentence.

FIG. 9 is a diagram illustrating an embodiment of a screen for displaying a search result output after a searcher searches for document data. FIG.

FIG. 10 is a diagram illustrating an embodiment of a screen for displaying a search result output after a searcher searches for an expected search sentence. FIG.

FIG. 11 is a diagram illustrating an embodiment of a screen on which an input result is output when data is input to a search database. FIG.

Explanation of symbols on the main parts of the drawings

10: Document address input section 11: Document name input section

12: Document description input unit 13: expected search input unit

21: unique number field 22, 23, 24: expected search field

25: Document Address Field 26: Document Name Field

27: Document description field 28: Similarity field

31: Search sentence input unit 32: Search button for expected search

33: search button 34: prediction search display unit

In order to achieve the above object of the present invention, a document data search database for document data search according to the present invention is

A document data basic information input step of inputting basic information about document data;

An expected search sentence input step of inputting an expected search sentence in a sentence form composed of a plurality of parts by extracting a search sentence expected in the corresponding document data input in the document data basic information input step;

A data conversion step of converting the data input in the document data basic information input step and the prediction search statement input step into an internal data format;

Opening a search database to open a search database after the data conversion step;

A search database input step of inputting the data converted in the data conversion step into a search database in a record form in accordance with each field of the search database;

And closing the search database.

Document data basic information input step is to input basic data necessary to build a search database for document data search, and to acquire basic information about document data and input basic information about document data. Basic information about documentation includes the name of the documentation, a brief description of the documentation, and information about the address where the documentation is located. In the case of document data constructed on the Internet, the document address refers to the Internet site address (URL, Uniform Resource Location). In the case of document data constructed on the intranet of an enterprise, the address indicating the path of the document data is indicated. I speak. In addition to the basic information presented above, information deemed necessary for constructing a search database can be additionally input.

Predictive search input step is a searcher inputs a randomly selected prediction search that is expected to be a search at the time of search. Extracting the expected search sentence in inputting the expected search sentence is not a way of using a keyword or a combination of words that can represent the contents of the document data used in a conventional search engine. It combines the meanings that are expected to be intended at the time of the search into appropriate words, divides them into parts according to the meaning of the sentences, and creates a search sentence in sentence form consisting of a plurality of parts and extracts it as an expected search sentence. That is, the extraction of the expected search sentence does not extract a keyword in the form of a keyword representing the content of a conventional document-oriented document, but in a form representing the intention of the searcher-centered searcher. It is to be divided into. For example, the first part corresponding to 'when, where, and who' and the second part corresponding to 'what,' are extracted. Part of the sentence and a third part corresponding to 'how to do' to extract and enter the expected search sentence. In this case, since the searcher wants to obtain something desired, it is preferable to attach the word 'I want' to the ending of the third part so that the search sentence is in the form of a smooth sentence. Here, as described above, dividing the predicted search sentence into three parts, the first part, the second part, and the third part, may be added or subtracted according to the construction condition of the search database. In addition, it is not necessary to input a search word in all parts of the prediction search input unit, and the search word may be formed by partially inputting the search word only in necessary parts according to a condition that the searcher intends. Of course, the method of dividing the expected search into parts according to the meanings can be applied in various ways using the above-described technical idea. For example, in the case of extracting a search sentence in a sentence form, for example, when the content of the document data is about 'climbing', a conventional search engine uses 'climbing' or 'mountain' as a keyword. Fully reflects what a searcher wants to search for document data on mountain climbing, and wants to “convert mood” or “enhance health,” which searchers are expected to use when searching, even if they are not directly related to the content of document data. It is to extract and construct a sentence in the form of a sentence such as 'I want to make a living', 'I want to have a leisure life', 'I want to go hiking' or 'I want to go to a mountain'. According to the meaning, the predicted search sentence extracted as described above is divided into the first part, the second part, and the third part described above according to the meaning, and inputted in the corresponding part. When the search database is entered into the search database, the searcher does not have to think of a separate keyword at the time of the search. Instead, the searcher inputs the search sentence directly in the sentence form in the search sentence input section to search. In particular, even beginners who have little experience in searching for documents can easily search for documents. In addition, by inputting the search sentence into several parts according to the meaning of the sentence, there is an advantage that the search can be made easily by combining the respective parts of the predicted search text searched by the predicted search text described below to simplify the search. .

In the data conversion step, the data entered in the data input step is inputted into the search database, outputted from the search database, or when the similarity rate is obtained using the data. It is the stage of unification. For example, removing the left and right spaces of the input string in the process of inputting basic information of the document and the expected search sentence, or making the consecutive spaces in the string into a single space, or inputting document address on the internet. In the case of a site address (URL), input data, such as unifying the format of an Internet site address (URL) by removing the '/' at the end of the Internet site address (URL), is converted to fit a certain format. Say that.

The opening of the search database is a step of opening a search database in order to build the data entered in the data input step and converted through the data conversion step in the search database.

The search database input step is a step of constructing a search database by inputting the data input in the data input step and converted through the data conversion step into the search database. The converted data is inputted into the corresponding fields, each with a unique number, an expected search statement, a document address, a document name, a document description, and a similarity rate, to form a record. It is entered into the search database in the form. At this time, the predicted search sentence is composed of a plurality of parts such as the first part, the second part, and the third part, as described above, and thus, the predicted search sentence 1 field, the predicted search sentence 2 field, and the predicted search sentence 3 field, respectively. The corresponding part is divided into inputs. The similarity rate is a value obtained by a similarity rate calculation method that determines the degree of similarity between a search word input by a searcher and an expected search word constructed in a search database. As the basic information data and the expected search text data of document data are continuously input, the search database is constructed as the number of records continues to increase as the records are continuously entered into the corresponding fields in the search database. It is going to be.

The closing of the search database is the step of closing and closing the search database in order to finish entering the search database when the search database is completed.

The closing of the search database may further include an input result output step of outputting an input result in order to confirm whether an input content is correctly inputted so that the search database is constructed.

In the present invention, a searcher searches for document data using a search database databased by a document data search database method for searching document data.

A search sentence input step of the searcher inputting a search sentence in a sentence form composed of a plurality of parts divided into parts according to the meaning of the search sentence input unit;

A data conversion step of converting the search statement data input in the search statement input step into an internal data format;

Opening a search database to open a search database;

Reads each record stored in the search database, calculates the similarity rate between the predicted search statement contents of each record and the search statement entered by the searcher, and enters the calculated similarity rate into the similarity field of each record. A similarity calculation step of updating a;

A record set construction step of constructing a recordset by selecting only records in which the similarity rate of the similarity rate field is greater than 0 in each record in which the similarity rate calculated in the similarity rate calculation step is input;

A recordset outputting step of outputting the contents of each record of the recordset configured in the recordset configuring step to a screen;

And closing the search database.

In the search sentence input step, the searcher who wants to search the document data inputs a search condition to find the document data that the user wants to search. The searcher divides the sentence that indicates what he intends to search in the search sentence input section into parts according to the meaning. Enter a search sentence in sentence form consisting of a plurality of parts. If the search sentence input unit is divided into three parts, a first part corresponding to 'when, where, and who', a second part corresponding to 'what', and a third part corresponding to 'how and' In this section, the searcher inputs what he wants. In this case, the search statement input unit may be added or subtracted according to the construction conditions of the search database, although it is not limited to the first part, the second part, and the third part. In order to improve the efficiency of the search by comparing the same parts when searching. In addition, the searcher may not necessarily input a search word in all parts of the search sentence input unit, but may partially input the search word only in the required part according to a condition intended by the searcher.

The data conversion step is a step of unifying the internal data format of the search database so that no abnormality occurs when a similarity rate is obtained using the search statement data input in the search statement input step.

The opening of the search database is a step of opening a search database in order to proceed with a search.

The similarity calculation step is already established in the search database, and each unique number field, expected search field, document address field, document name field, and document description are provided with the unique number, the expected search statement, the document address, the document name, and the document description as fields. Similarity calculation method that reads the records with the corresponding contents in the field one by one for use in document data search, and sets the similarity rate with the search sentence entered by the searcher in the expected search sentence field of each read record. And updating each record by inputting the calculated similarity rate into the similarity rate field of each record. At this time, the method of calculating the similarity rate can be configured in various ways. An example of a method of calculating the similarity rate will be described below. In this case, the similarity calculation step refers to a process of calculating the similarity rate between the predicted search sentence in the predicted search sentence field for all records and the search input entered by the searcher by reading all the records constructed in the search database step by step. The updated similarity rate is entered in the similarity rate field of each record of the search database to update each record. In other words, it reads one record to calculate the similarity rate and updates the record, and then reads the next record to calculate the similarity rate and updates the record to repeat the record reading, similarity calculation and record update for all records. do. When the record update is completed by reading the entire record and calculating the similarity rate, the next step is to construct the record set.

5 is a flowchart illustrating an embodiment of a process in which a similarity rate calculation step is performed. As shown in FIG. 5, a first step of moving a position where a record is read from a search DB to a first line of the record, a second step of determining whether the position of the record moved in the first step is a last row, and the first step If it is determined in step 2 that it is not the last line, a third step of reading a record, a fourth step of calculating a similarity rate between the predicted search statement of the record read in the third step and the search statement input by the searcher; A similarity calculation including a fifth step of updating a record by inputting the similarity rate calculated in step 4 into a similarity rate field of the current record, and a sixth step of moving to a record position of the next line when updating is completed in the fifth step. The steps are in progress. If the position of the record is determined to be the last line in the second step, the reading of the record is completed and the record set construction step is performed.

In the recordset configuration step, only records having a similarity rate greater than zero among each updated record are selected and configured into a recordset. If the calculated similarity rate is less than or equal to 0, it is not a similar search sentence entered by the searcher, so the document data of the record is not the document data that the searcher wants, so it is not necessary to provide the record to the searcher. Only cases with greater than zero are useful to the searcher, so select only records with similarity greater than zero to construct the recordset.

The record set output step is a step of outputting the contents of each record of the record set configured in the record set configuration step so that a searcher can view the screen. The searcher can check the list of relevant document data according to the search statement entered by the user by checking the recordset displayed on the screen, and the contents provided to the searcher through the output of the recordset are inputted by the searcher. Includes search, expected search, similarity rate, document name, document address, and document description. The searcher can find the document data that best matches the intended search purpose by checking the similarity rate between the search text entered by the searcher and the expected search text stored in the search database. At this time, in the step of outputting the configured recordset on the screen, the similarity value of each record in the recordset is compared, and the records having the similarity value are sorted in order of the smallest records and output to the searcher. It is desirable to be able to check the document data.

The closing of the search database is a step of closing a search database by completing a search process for a search statement input by a searcher through the record set output step.

In searching the document data using a search database constructed by a document data search database method for searching document data according to the present invention, the searcher can search the expected search text input and stored in the search database in advance. It is possible to further increase the convenience and efficiency of the searcher's search by using the predicted search text that is input and stored in the document data search.

As described above, a searcher searches for an expected search sentence using a search database databased by a document data search database method for searching document data of the present invention.

A search sentence input step of the searcher inputting a search sentence in a sentence form composed of a plurality of parts divided into parts according to the meaning of the search sentence input unit;

A data conversion step of converting the search statement data input in the search statement input step into an internal data format;

Opening a search database to open a search database;

Reads each record stored in the search database, calculates the similarity rate between the predicted search statement contents of each record and the search statement entered by the searcher, and enters the calculated similarity rate into the similarity field of each record. A similarity rate calculating step of updating a;

A record set construction step of constructing a recordset by selecting only records in which the similarity rate of the similarity rate field of each record updated in the similarity rate calculation step is greater than zero;

An expected search statement outputting step of outputting an expected search statement in the predicted search statement display unit of the record set configured in the record set configuration step;

And closing the search database.

The searcher input step, data conversion step, search database open step, similarity rate calculation step, recordset configuration step, and search database closing step may include the searcher described above. This is the same function as each step included in the method of searching the document using the search statement.

The predicted search text searched in the predicted search text output step is displayed in the predicted search text display unit. The predicted search text is output in a sentence form composed of a plurality of parts divided into parts according to meanings. As in the above-described example, the first part corresponding to 'when, where, and who', the second part corresponding to 'what', and the third part corresponding to 'how' It can be output to the searcher in the form of sentences consisting of. If there are multiple predictions for records with similarity greater than zero, multiple predictions are displayed in the prediction display. The predictive search output displayed in the predictive search display section allows the search text to be linked to the search text input section when the searcher clicks on a predicted search text in the predictive search display section so as to save typing effort in the search. It is desirable to.

Hereinafter, with reference to the accompanying drawings will be described in detail a preferred embodiment of the present invention.

1 is a diagram showing the flow of data of the present invention. As shown in FIG. 1, data of a search database is input through an input process, and a search database is constructed. A searcher who wants to search document data searches for document data through a search process using the established search database. In search of document data, it can be used to search the document data by searching the expected search text stored in the search database through the search text search process.

2 is a flowchart illustrating a document data retrieval database process. As shown in FIG. 2, a person who intends to make a document data search database for document data search first obtains basic information on document data to be searched and extracts an expected search statement for the document data. Extracting the expected search sentence is the biggest feature of the present invention, and extracts a search sentence that can be used when the searcher wants to find the corresponding document data from the position of the searcher to make the expected search sentence in sentence form. At this time, since the extracted search sentence is a sentence form, searchers can easily use the search. Enter the basic information of the acquired document and the extracted prediction. The data input in the data input step is converted into an internal data format, and then the search database is opened and entered into the search database in a record form for each field of the search database. It will end by closing.

3 is a flowchart illustrating a process of a searcher searching for document data through a search database. As shown in FIG. 3, the searcher inputs a search sentence in a sentence form composed of a plurality of parts divided into parts according to the meaning of the search sentence input unit. Next, the search statement data input by the searcher is converted into an internal data format. . Next, open the search database to search the document data, read each record stored in the search database, calculate the similarity rate between the predicted search text in the predicted search field of each record and the search text entered by the searcher for each record. Update the record by entering the calculated similarity rate into the similarity field of each record. Next, in each of the records updated in the similarity rate calculation step, only a record having a similarity rate of greater than zero in the similarity rate field is selected to constitute a recordset. The recordset is then displayed on the screen so that the searcher can see the contents of each record in the configured recordset. At this time, the recordset displayed on the screen is compared by the similarity rate of each record and is sorted in order from the record with the largest similarity value to the smallest record.

4 is a flowchart illustrating a process of a searcher searching for an expected search sentence through a search database. As shown in FIG. 4, the searcher inputs a search sentence in a sentence form composed of a plurality of parts divided into parts according to the meaning of the search sentence input unit. The search statement data entered by the searcher is converted into internal data format. The search database is then opened, and each record stored in the search database is read, and for each record, the similarity rate is calculated between the contents of the expected search statement field of each record and the search statement entered by the searcher. The record is updated as it is entered in the similarity field. Next, only records in which the similarity rate in the similarity field of each updated record is greater than zero are selected to form a recordset. In the next configured record set, the prediction search statement is displayed in the prediction search statement display section of the screen, and the search database is closed. At this time, the predicted search output displayed in the predicted search display unit is linked to move the search sentence to the search sentence input unit when the searcher clicks on a predicted search sentence of the predicted search display unit in order to save typing effort in the search. have.

6 is a view illustrating an embodiment of a screen for inputting basic document information and an expected search text in order to construct a search database. As shown in FIG. 6, a document address input unit 10 for inputting a document address, a document name input unit 11 for inputting a document name, a document description input unit 12 for inputting a document description, and an expected search statement An expected search text input unit 13 is provided. The predictive search input unit is divided into parts according to meanings, and is composed of a plurality of parts. As shown in FIG. 5, a first part corresponding to 'when, where, and who', and a second part corresponding to 'what' and In addition, it is possible to input the expected search sentence in the sentence form which is divided into 3 parts corresponding to 'how and'.

FIG. 7 is a diagram illustrating an embodiment of a screen that displays input results so that basic document information and expected search statements can be inputted into a search database. As shown in FIG. 7, when a record has a unique number, an expected search statement divided into three parts, a document data address, a document name, a document description, and similarity data, each unique number field is provided in the search database. (21), Predictive Search 1 field (22), Predictive Search 2 field (23), Predicted Search 3 field (24), Document Data Address field (25), Document Name field (26), Document Description field ( 27), the similarity rate field 28 is configured so that the corresponding data is input and output.

FIG. 8 is a diagram illustrating an embodiment of a screen in which a searcher input unit for inputting a search sentence is provided. As shown in FIG. 8, the search sentence input unit 31 is divided into parts according to meanings, and thus, the first part corresponding to 'when, where, and who', as shown in FIG. The search sentence can be entered in a sentence form that is divided into a second portion corresponding to 'and a third portion corresponding to' how to do '. At this time, when the search sentence input unit 31, respectively, after entering the search sentence and click on the search sentence search button 32, compared with the input search sentence predicted sentence with a similarity ratio greater than 0 is the expected search sentence display unit 34 When the search text input unit 31 inputs the search text and clicks the search button 33, the document data having a similarity rate greater than 0 compared to the search text compares the similarity rate value of each record. According to the size of the rate value, the similarity value is output in the order of the smallest records from the largest records.

FIG. 9 is a diagram illustrating an embodiment of a screen for displaying a search result output after a searcher searches for document data. As shown in FIG. 9, the search results, the document name, the document data address, and the document description inputted at the time of constructing the search database are outputted in the search result, and the similarity ratio between the search word input by the searcher and the expected search word is calculated. The value is output to help the searcher search.

FIG. 10 is a diagram illustrating an embodiment of a screen for displaying a search result output after a searcher searches for an expected search sentence. As shown in FIG. 10, when each search sentence is input to the search sentence input unit 31 and the search sentence search button 32 is clicked, the predicted search having a similarity rate greater than 0 compared to the search sentence input through the search database The statement is output to the expected search statement display unit 34.

FIG. 11 is a diagram illustrating an embodiment of a screen on which an input result is output when data is input to a search database. As shown in FIG. 11, the input result is outputted so that basic information of the document to be input at the time of constructing the search database and the expected search statement are correctly input.

In addition, as described above, the method of calculating the similarity rate between the predicted search text and the search text can be implemented by various methods. Hereinafter, an embodiment of a method of calculating the similarity rate will be described.

When sentence A is 'useful and easy internet search' and sentence B is 'useful and interesting search', the method of calculating the similarity rate between sentence A and sentence B is 'useful' and 'easy' Calculate ',' 'Internet', 'Search' as one word, and calculate 'Useful', 'Funny' and 'Search' as one word in sentence B. Calculate

The number of words in sentence A is four, and the number of words in sentence B is three, so the total number of words is seven. The number of words in the sentence to be compared is as follows.

The number of words that are substrings of sentence A in sentence B is two:

The number of words in sentence B that are substrings of sentence A is 1:

Thus, the total number of match words is three.

Therefore, the similarity rate between sentence A and sentence B is

Becomes

The similarity rate calculation method described above is just one embodiment, and the similarity rate calculation may be calculated by mathematical calculations using various methods.

In the present invention described above, the document data search database and the document data search method using the same are provided with search sites on the Internet based on the technical idea of the present invention described above. It can be implemented in the form that can be searched, and also in the form that a searcher can input the search text and receive the search result through other wired / wireless communication means, and can be implemented by forming an independent document data search program. .

Based on the technical idea of the present invention, a search database for searching document data on an Internet site constructed on the Internet and a search method using the same can be implemented, and for searching document data on an intranet established in an enterprise or a group. It turns out that it is possible to implement a search database and a search method using the same.

According to the present invention, by searching the document data and searching the document data using the same, the searcher sets the search term in searching the document data, eliminates the inconvenience of configuring the search formula, and intends the general sentence form. This is an advantage that you can easily search the document data by searching the document data by naturally entering the search sentence of the search statement, and the search statement is searched only in the same meaning object by dividing the search statement according to the meaning in sentence form. If the searcher doesn't know in detail which search terms to search, the user can simply enter the searcher's intention and then search for the predicted search sentence to obtain the prediction search similar to the searcher's intention. Is there.

Claims (7)

As a document data search database for searching document data, Inputting document data basic information for inputting basic information about document data; An expected search sentence input step of inputting an expected search sentence in a sentence form composed of a plurality of parts by extracting a search sentence expected in the corresponding document data input in the document data basic information input step; A data conversion step of converting the data input in the document data basic information input step and the prediction search statement input step into an internal data format; And And a search database input step of inputting the data converted in the data conversion step into a search database in a record form corresponding to each field of the search database. How to document search database to make it searchable. The method according to claim 1, further comprising an input result output step of outputting an input result in order to confirm whether the input content is correctly input and built into the search database in the document data search database forming method. How to document search database. The method of claim 1, wherein the document data basic information input in the document data basic information input step includes at least a name of document data, a description of document data, and an address of document data. The method of claim 1, wherein the predicted search sentence input in the predicted search sentence input step includes a first part corresponding to 'when, where, and who', a second part corresponding to 'what', and 'how' Method for document search database, characterized in that divided into three parts corresponding to the. The record of claim 1, wherein the record of the search database generated in the search database input step includes at least a unique number field, an expected search sentence field, a document address field, a document name field, a document description field, and a similarity rate field. Document data retrieval database method characterized in that the corresponding data is input. As a method of searching document data using a search database databased by a document data search database method that enables searching of document data using an expected search statement, A search sentence input step of the searcher inputting a search sentence in a sentence form composed of a plurality of parts divided into parts according to the meaning of the search sentence input unit; A data conversion step of converting the search statement data input in the search statement input step into an internal data format; Reads each record stored in the search database, calculates the similarity rate between the predicted search statement contents of each record and the search statement entered by the searcher, and enters the calculated similarity rate into the similarity field of each record. A similarity rate calculating step of updating a; A record set construction step of constructing a recordset by selecting only records in which the similarity rate of the similarity rate field of each record updated in the similarity rate calculation step is greater than zero; And And a recordset output step of outputting the contents of each record of the recordset configured in the recordset configuration step to the screen. As a method of searching document data using a search database databased by a document data search database method that enables searching of document data using an expected search statement, A search sentence input step of the searcher inputting a search sentence in a sentence form composed of a plurality of parts divided into parts according to the meaning of the search sentence input unit; A data conversion step of converting the search statement data input in the search statement input step into an internal data format; Reads each record stored in the search database, calculates the similarity rate between the predicted search statement contents of each record and the search statement entered by the searcher, and enters the calculated similarity rate into the similarity field of each record. Updating the record; A record set construction step of constructing a recordset by selecting only records in which the similarity rate of the similarity rate field of each record updated in the record updating step is greater than zero; And Document prediction data search to search the document data, characterized in that it comprises a step of outputting the expected search statement outputting the expected search statement in the predicted search sentence display unit on the screen set configured in the record set configuration step Search method.