JP4516103B2 - Server apparatus and computer program - Google Patents

Description

  The present invention relates to a server device and a computer program.

  In recent years, the Internet has spread rapidly, and a large amount of information is available via the Internet. Search engines are the most effective means for efficiently handling such an enormous amount of information.

  Search engines include a search engine that searches a large number of unspecified information on the Internet using a search search word, and a search engine that searches information in a specific website according to the search search word.

Also, since search users can find commonality even among different users, a technology called a suggest function is also provided in recent years in which search sites present search word candidates to reduce the user's input burden. (See Patent Document 1)
JP 2002-324077 A

  However, in the above suggestion functions, there are many that present search terms that the user has entered in the past, search terms that have been used in the past on the search site, and suggest search terms candidates based on the internal dictionary. It is not always possible to accurately present a combination of words that are uniquely used in a specific website or highly related words that appear in the same page.

  Therefore, an object of the present invention is to provide a technique for presenting a search term actually used in a website as a candidate when performing a search in the website.

The present invention for solving the above problems is as follows.
In response to a search instruction from a client device that accesses a website published on the Internet, the server device performs a search in the website,
A first database for performing a full text search of information of web pages constituting the website;
A second database in which words contained in the web page are registered;
A receiving means for receiving an input word input for searching the website in the client device each time a new character is input;
Search means for searching the second database for a word that matches the received input word and determining a search word candidate;
Transmission means for transmitting the search word candidate determined by the search means to the client as a search word;
Search means for performing a full-text search using the first database based on the search terms received from the client device;
When the input word received by the receiving means includes a delimiter,
The search means includes
Searching the second database for a first word that matches the first input word immediately before the delimiter;
Searching for the first web page containing the searched first word using the first database;
Extracting a first neighboring word existing around the first word in the searched first web page;
The first word and the first neighboring word are determined as the search word candidates.

  ADVANTAGE OF THE INVENTION According to this invention, when performing the search in a website, the technique for showing the search term actually used in the site as a candidate can be provided.

  Embodiments of the present invention will be described below with reference to the accompanying drawings. However, the constituent elements described in this embodiment are merely examples, and are not intended to limit the scope of the present invention only to them.

<System configuration>
FIG. 1 is a block diagram showing the overall configuration of a search system in a website corresponding to this embodiment. Clients 101a, 101b, and 101c (hereinafter collectively referred to as “client 101”) that perform a search in the website by the user's operation, provide the website information, and the server 103 that performs the search in the website. Are connected to the Internet 102. In addition, a site information database 104, a full text search database 105, and a search term database 106 are connected to the server 103.

  The client 101 is an information processing apparatus that can be operated by a user and browse website information published by the server 103 on the Internet, and includes a notebook personal computer, a desktop personal computer, a portable information terminal, a mobile phone, and the like. It is assumed that so-called Internet browser software is installed in the client 101.

  The client 101 is connected to the Internet 102 by a LAN which is an access means including a network cable such as Ethernet (registered trademark). Note that the access means to the network 102 is not limited to an Ethernet (registered trademark) cable, and may be constituted by wireless communication means such as a wireless LAN. It is assumed that a server 103 to be described later includes similar access means.

  The Internet 102 is a network in which networks all over the world are connected to each other, but may be a network that can be connected only within a specific organization, such as an intranet.

  The server 103 publishes on the Internet a website that provides information on a specific company, store, or individual to the user of the client 101, and at the same time, an information processing apparatus that provides a search service for web pages in the website It is. In the present embodiment, for the sake of simplicity of explanation, a case has been described in which the server 103 performs both site publication and site search, but the configuration of the server 103 is not limited to this. That is, a server for publishing a site and a server for searching in a site can be provided separately. Note that a website refers to a group of web pages published as a group. Websites include, for example, corporate and personal websites.

  The server 103 manages information stored in the site information database 104, the full text search database 105, and the search term database 106. The server 103 is connected to the site information database 104, the full-text search database 105, and the search term database 106 by, for example, LAN.

  The site information database 104, the full-text search database 105, and the search term database 106 are information processing apparatuses in which predetermined database software is installed, and manage various data.

  The site information database 104 manages information on web pages for publishing websites on the Internet 102. The web page includes information on the website owner and products provided by the owner. For example, in the case where the website is a corporate website, information on products manufactured and sold by the corporation and information on all services provided by the corporation are included in addition to the company outline and information for investors.

  The full-text search database 105 is a database in which words and page meta information (URL, title, search order, etc.) are linked based on the web page stored in the site information database 104. The server 103 performs a search using the full-text search database 105 using the search term specified by the user received from the client 101.

  The search word database 106 extracts words contained in web pages constituting the site from the search information stored in the full-text search database 105 and manages them in a table. The server 103 inputs input words from the client 101. When received, search word candidates are provided according to the registered contents of the table.

  In this specification, for the sake of convenience, the server 103, the site information database 104, the full-text search database 105, and the search word database 106 are described as being realized by physically independent information processing apparatuses. The embodiment of the invention is not limited to this. For example, these may be realized by a single information processing apparatus. On the other hand, each device such as the server 103 may be configured redundantly or distributedly by a plurality of information processing devices. The site information database 104, the full-text search database 105, and the search word database 106 are described as being connected to the server 103 via a LAN or the like. For example, the site information database 104, the full-text search database 105, and the search term database 106 can communicate with the server 103 via the Internet 102 or an unshown intranet. May take any form.

<Configuration of information processing apparatus>
Next, an outline of the information processing apparatus constituting the website search system corresponding to the present embodiment will be described. FIG. 2 is a block diagram illustrating an example of the hardware configuration of the client 101. The site information database 104, the full-text search database 105, and the search word database 106 as the information processing apparatus described above may be configured in the same or equivalent hardware configuration.

  In FIG. 2, a CPU 200 executes an application program, an operating system (OS), a control program, and the like stored in a hard disk device (hereinafter referred to as HD) 205, and stores information, files, etc. necessary for executing the program in a RAM 202. Control to temporarily store.

  The ROM 201 stores various data such as a program such as a basic I / O program, font data used for document processing, and template data. The RAM 202 temporarily stores various data and functions as a main memory, work area, and the like of the CPU 200.

  The external storage drive 203 is an external storage drive for realizing access to a recording medium, and can load a program or the like stored in the medium (recording medium) 204 into the computer system. The media 204 is, for example, a floppy (registered trademark) disk (FD), CD-ROM, CD-R, CD-RW, PC card, DVD, Blu-ray, IC memory card, MO, memory stick, or the like. can do.

  In this embodiment, the external storage device 205 uses an HD (hard disk) that functions as a large-capacity memory. The HD 205 stores application programs, OS, control programs, related programs, and the like. A nonvolatile storage device such as a flash (registered trademark) memory may be used instead of the hard disk.

  The instruction input device 206 corresponds to a keyboard, a pointing device (such as a mouse), a touch panel, or the like. Using the instruction input device 206, the user instructs the client 101 to input a command or the like for controlling the device. The display 207 displays a command input from the instruction input device 206, a response output of the client 101 in response thereto, and the like. A system bus 209 manages the flow of data in the information processing apparatus. An interface (hereinafter referred to as I / F) 208 plays a role of mediating data exchange with an external device.

  In addition, it can also be comprised as an alternative of a hardware apparatus with the software which implement | achieves a function equivalent to the above each apparatus.

  In the present embodiment, an example is shown in which the program and related data according to the present embodiment are directly loaded from the medium 204 to the RAM 202 and executed. However, every time the program according to the present embodiment is operated, the program is already executed. May be loaded into the RAM 202 from the HD 205 in which is installed. It is also possible to record the program according to the present embodiment in the ROM 201, configure it as a part of the memory map, and execute it directly by the CPU 200.

  FIG. 3 is a block diagram illustrating a device configuration of the server 103. Functions and applications of the CPU 200, ROM 201, RAM 202, external storage drive 203, media 204, HD 205, instruction input device 206, display 207, I / F 208, system bus 209, and their relationship will be described with reference to FIG. Similar or equivalent to Note that in FIG. 3, database 310 is connected to system bus 209. Here, the database 310 is a generic term for the site information database 104, the full text search database 105, and the search word database 106.

  Next, a search term table stored in the search term database 106 corresponding to the embodiment of the invention will be described. In the present embodiment, a word table, a kana conversion table, and a romaji conversion table are provided as search term tables.

  The word table registers words that can be search word candidates as candidate words. In the present embodiment, words registered as candidate words can be written using kanji, hiragana, katakana, alphabets, symbols, numbers, and synthesized words. The symbol refers to a mark with a specific meaning such as “!”, “%”, “¥” or the like. Symbols also include arithmetic symbols such as “+”, “−”, “×”, “÷”. The numbers include Roman numerals and the like in addition to the arithmetic numbers 1 to 9. Therefore, a mathematical expression composed of a combination of numbers and symbols can be registered as a word. In addition, a composite word is a combination of specific characters, or a character given a different reading or meaning different from the reading or meaning of the character. For example, the inequality sign “<” gives the reading of “Hi” in Hiragana, “Lena” in Katakana gives the reading of “K” in Hiragana, and the meaning of “present” in “Omi” in Kanji. Or give.

  FIG. 4 is a diagram illustrating an example of a word table. In FIG. 4, the word table 400 includes candidate words 401 and points 402. In the candidate word 401, a word that can be a search word when performing a search in a website is registered. The words registered as the candidate words 401 are only words included in the web page registered in the site information database 104. As the point 402, a point given to the candidate word 401 is registered. The point is given to each word according to the number of appearances, the appearance position, and the display size in the website. In addition, in consideration of the distribution in the web page, points may be reduced for words with a high distribution rate. The higher the word, the more important the word on the website.

  Next, the kana conversion table is a table for registering kana notation of the candidate word 401 of the word table 400. The kana conversion table corresponding to the present embodiment can register hiragana notation of words (in particular, kanji, katakana, alphabets, symbols, numbers, and composite words) registered in the word table 400. FIG. 5 is a diagram illustrating an example of a kana conversion table. In FIG. 5, candidate words 501 and kana candidates 502 are registered in the kana conversion table 500. Candidate words registered in the word table 400 are registered in the candidate word 501. In the kana candidate 502, the kana notation of each candidate of the candidate word 501 is registered. For example, “hanbai” is registered in the kana candidate corresponding to the candidate “sales”.

  Next, the romaji conversion table is a table for registering romaji notation of kana candidates registered in the kana conversion table. FIG. 6 is a diagram illustrating an example of a Roman character conversion table. In FIG. 6, kana candidates 601 and romaji candidates 602 are registered in the romaji conversion table 600. As the kana candidate 601, a kana candidate registered in the kana conversion table 500 is registered. In the Romaji candidate 602, the Roman character notation of each candidate of the kana candidate 601 is registered. For example, “opuson” is registered for “Opuson” and “hanbai” is registered for “hanbai”.

  Note that the embodiment of the invention is not limited to the configuration having the kana conversion table 500 and the romaji conversion table 600 in addition to the word table 400, but may have only the kana conversion table 500. Alternatively, the Roman alphabet notation of the words registered in the word table 400 may be registered in the Roman alphabet conversion table 600, and only the Roman alphabet conversion table 600 may be included.

  Next, based on the configuration shown in FIG. 1 to FIG. 6, the details of the website search process executed by the website search system of the present invention will be described.

  FIGS. 7a to 7g are state transition diagrams showing an example of the search processing in the website corresponding to the embodiment of the invention according to the description method of SDL (Specification and Description Language). The processing corresponding to FIGS. 7a to 7g is realized by executing the processing programs held by the client 101 and the server 103, respectively. Further, communication between the client 101 and the server 103 can be realized using a communication function of a web browser executed by the client 101 or a communication function of a plug-in (extension program) of the web browser. For example, it can be realized according to an HTTP protocol using Java (registered trademark) Script. In addition, Flash or the like may be used, or communication may be performed according to another protocol that is not HTTP.

  First, FIG. 7 a is a diagram illustrating state transition between the client 101 and the server 103. In FIG. 7a, the client 101 takes the state from step S710 to step S760, and the processing shown in FIGS. 7b to 7f is performed in each state. In addition, the server 103 is in a state of waiting for a request from the client 101, and executes processing as shown in FIG. 7g described later in response to a request received from the client 101.

  First, the client 101 is in an initial state in step S710. In this initial state, the search screen is not displayed on the display 207 of the client 101, and a search screen display request is transmitted to the server 103 in step S711 as shown in FIG. 7b. Thereafter, the process proceeds to the search screen display waiting state in step S720.

  In the search screen display state in step S720, as shown in FIG. 7c, the client 101 receives the search screen information from the server 103 in step S721, displays the search screen on the display 207 in step S722, and the search screen in step S730. Transition to the display state.

  An example of the search screen is as shown in FIG. In FIG. 9, the search screen 900 includes a search word input area 901 and a search execution button 902. When the characters constituting the search word are input in the search word input area 901 as kanji, hiragana, katakana, alphabet, symbol, number, or composite word, the search word candidate display area 903 is searched according to the procedure described below. Word candidates are displayed. The user can select an appropriate search word from the search word candidates in the search word candidate display area 903 and input the search word.

  In the search screen display state in step S730, as shown in FIG. 7d, three types of events are executed according to the input content from the user.

  The first is a character input event in step S731, where a kanji, hiragana, katakana, alphabet, symbol, number, or composite word is input to the search word input area 901 in FIG. When this event occurs, the input word is transmitted from the client 101 to the server 103 in step S732. Thereafter, the process shifts to a waiting state for receiving search word candidates from the server 103 in step S740.

  Note that a character input detection program for performing processing for detecting character input and transmitting it to the server 103 is included in the search screen information received from the server 103 in step S721, and the client 101 receives the search screen information. A character input detection program can be acquired each time. This character input detection program can be constituted by, for example, Java (registered trademark) Script, Flash.

  Next, the second is a search word candidate selection event in step S733, and a desired search word candidate is selected from the search word candidates displayed in the search word candidate display area 903 of FIG. The selection operation is, for example, by mouse click or Enter key operation. When a search word candidate is selected, a delimiter is added to the search word and transmitted as an input word from the client 101 to the server 103 in step S734. Thereafter, the process shifts to a waiting state for receiving search word candidates from the server 103 in step S740. Note that the delimiter includes a space (blank), an AND symbol (+), and the like, but is not limited to those illustrated here, and may be any symbol or character used to separate two words.

  Furthermore, the third is a search execution event in step S735, which accepts a search execution instruction from a user based on a predetermined operation. The search execution instruction can be received based on, for example, a click operation of the search button 902 using the mouse of the instruction input device 206. In addition, it may be based on the operation of the Enter key of the keyboard, or based on an arbitrary operation input using Java (registered trademark) Script or the like. In a succeeding step S736, a search execution instruction specifying a search word is transmitted from the client 101 to the server 103. Thereafter, the process proceeds to the search result reception waiting state in step S750.

  In the search word candidate waiting state in step S740, as shown in FIG. 7E, the search word candidate is received from the server 103 in step S741. Subsequently, the search word candidate received in step S742 is displayed in the search word candidate display area 903 of the search screen 900 in FIG. 9, and the process proceeds to the search screen display state in step S730.

  In the search word result reception waiting state in step S750, the search result is received from the server 103 in step S751, as shown in FIG. Subsequently, the search result received in step S752 is displayed on the display 207, and the process proceeds to the search result display state in step S760. In the search result display state in step S760, it is possible to accept specification of a page to be accessed from the user based on the search result.

  Next, processing on the server 103 side will be described. In step S770 of FIG. 7g, the server 103 is initially in a request waiting state. In this request waiting state, the server 103 receives any of the search screen display request transmitted from the client 101 in step S711, the input word transmitted in step S732 or S734, and the search execution instruction transmitted in step S736. To do.

  When a search screen display request is received in step S771, search screen information is transmitted to the client 101 in step S772. Thereafter, the process returns to the request waiting state in step S770.

  Next, when an input word is received in step S773, the process proceeds to step S774 to search for a search word candidate, and in step S775, the searched search word candidate is transmitted to the client 101. Thereafter, the process returns to the request waiting state in step S770. The search processing for search word candidates will be described later with reference to the flowchart of FIG.

  Furthermore, when a search execution instruction is received in step S776, the process proceeds to step S777 to execute search processing. This search process is performed using the full-text search database 105 based on the search term included in the search execution instruction. In the subsequent step S778, the search result is transmitted to the client 101, and the process returns to the request waiting state in step S770.

  Next, details of the search processing for a search word candidate in step S774 will be described with reference to FIG.

  In step S801, it is determined whether or not a delimiter is included in the input word received in step S705. Note that no character may be input after a space or the like.

  If no delimiter is included (“NO” in step S801), the process proceeds to step S803. On the other hand, when a delimiter is included (“YES” in step S801), the process proceeds to step S802.

  In step S802, the character string immediately before the delimiter is selected as an input word to be processed, and the process proceeds to step S803. In step S803, the word table 400 is searched based on the input word, and a candidate word that matches forward is searched. For example, if “op” is an input word, “opson”, “opson printer”, “opson direct”, “opson square”, “opson service”, etc. are searched.

  Next, in step S804, it is determined whether the input word is composed of hiragana or alphabet. If only one of the kana conversion table 500 and the romaji conversion table 600 exists in the search word database 106, it is determined only whether it is hiragana or alphabet according to the corresponding table. If it is composed of hiragana or alphabet (“YES” in step S804), the process proceeds to step S805. If it is composed of elements other than hiragana or alphabet (“NO” in step S804), the process proceeds to step S806.

  In step S805, if the input word is hiragana, the kana conversion table 500 is used to search for a candidate word that matches forward. For example, when “han” is an input word, “sale”, “sales corporation”, “at the time of sale”, “under sale”, “range”, etc. are searched. If the input word is alphabetic characters, the roman character conversion table 600 is used to search for a kana candidate that matches forward. For example, if “op” is an input word, search for “opson”, “opsompurinta”, “opsodairikuto”, “opsonsukueia”, “opsonsabis”, etc. Is done. The searched kana candidates are converted into candidate words using the kana conversion table 500. Thereafter, the process proceeds to step S806. In the case where the kana conversion table 500 does not exist in the search word database 106 and only the Roman character conversion table 600 exists, the Roman word notation is directly converted into a candidate word.

  In step S806, search term candidates are determined. Specifically, a predetermined number of candidate words with high points can be selected from the candidate words searched in steps S803 and S805, and determined as search word candidates.

  Next, in step S807, it is determined whether or not a delimiter is included in the input word in the determination in step S801. If a delimiter is included (“YES” in step S807), the process proceeds to step S808. On the other hand, if no delimiter is included (“NO” in step S807), the search word determination process ends, and the process proceeds to step S707 in FIG.

  In step S808, the web page including the search word candidate determined in step S806 (hereinafter referred to as “first search word candidate”) is searched using the full-text search database 105. In subsequent step S809, words (neighboring words) located around the first search word candidate are extracted from the web page detected by the search in step S808. “Located in the vicinity” can be defined as follows, for example. Located in the same sentence as the sentence including the first search word candidate, or further included in the sentences before and after the sentence, or before and after the first search word candidate, and within the predetermined number of characters from the first search word candidate A word.

  In subsequent step S810, a word that matches the character string immediately after the delimiter is selected from the peripheral words extracted in step S809, and the peripheral words are narrowed down. If there is no character string immediately after that, that is, if only a blank is input, this step can be omitted, and the subsequent processing can be performed for all the peripheral words obtained in step S809.

  In subsequent step S811, a distance Dw between each of the surrounding words obtained through the above steps and the first search word candidate is calculated. The distance Dw can be determined based on, for example, the same sentence as the sentence including the first search word candidate or the preceding or following sentence, the number of characters positioned between the surrounding word and the first search word candidate, or the like. .

  In subsequent step S812, the appearance frequency Fw of each neighboring word is calculated. The appearance frequency Fw can be based on, for example, the frequency (number of times) that the same surrounding word appears in the same page. Peripheral words with higher frequency can be considered as words that are more likely to be related to the first search word candidate in the page. However, even a word that appears frequently on the same page may be a general word on the site. Therefore, a value (Aw / Pal) obtained by dividing the total number Aw of the number of appearances of neighboring words in all pages (total number of pages is Pal) including the first search word candidate detected in step S808 by the total number of pages. Ask. If this value is larger than a predetermined threshold, the surrounding word paired with the first search word candidate may not be a word characteristic of a specific page, so the Fw point is lowered by a certain amount, or , It may be excluded from the peripheral word candidates.

  Next, in step S813, points are calculated for each peripheral word based on the distance Dw calculated in step S811 and the appearance frequency Fw calculated in step S812. The point calculation method is not particularly limited. For example, the sum or product of Dw and Fw may be used, or calculation may be performed after weighting. Further, only one of Dw and Fw may be used.

  In step S813, the points are calculated based on the distance and the appearance frequency. However, it is also possible to perform calculation in consideration of part-of-speech analysis and dependency of the first search word candidate. Further, for each combination of the first search word candidate and the surrounding word, the selection frequency selected in the past and the order given in advance (may be specified by the user) are further considered, or points based on these factors Can also be calculated. Specifically, a high point can be given to a combination with a high frequency selected in the past or a combination with a high ranking. Information on the selection frequency and rank can be stored in the search word database 106.

  Next, in step S814, surrounding words whose calculated points are greater than or equal to a predetermined value are determined. Note that this step may be omitted and all the peripheral words for which points have been calculated in step S813 may be selected. In the subsequent step S815, the determined peripheral word and the first search word candidate are determined as new search word candidates (second search word candidates). That is, the combination of the first search word candidate in step S806 and the surrounding words in step S814 becomes the second search word candidate. At that time, the point of the second search word candidate is the point obtained in step S813. Thereafter, the search word candidate determination process is terminated, and the process proceeds to step S707 in FIG.

  If the input word contains three or more words, that is, if two or more delimiters are included, the processing after step S808 is repeated for the character string immediately after each delimiter to further narrow down. The third and fourth search word candidates can be determined.

  As described above, according to the present embodiment, a word extracted in advance from a web page in a website can be presented to a user as a search word candidate when performing a search. A search can be performed using existing words as search terms.

  In addition, when a user intends to input a plurality of words as search terms, a combination of words that are actually relevant on a page in the website can be presented as candidates, so that a more accurate search is possible. Become.

  In particular, in the present invention, when determining search word candidates for the second word and thereafter, candidates are determined by real-time calculation without preparing a table in advance. Thereby, compared with the case where the combination candidate is registered in the database in advance, the capacity cost of the database can be greatly reduced.

  Further, even if the search word candidates for the second word and thereafter are not input, a search word candidate highly relevant to the first word is presented only by inputting a delimiter such as a space after inputting the first word. Therefore, even if the user does not think about the second word by himself / herself, he / she can efficiently perform a search by selecting a search word that will correspond to the page he / she is looking for from the presentation candidates. In this case, the present invention does not present a pair of words that have been input in the past based on the history as a candidate, but particularly presents a pair of words that are recognized to be relevant on an actual page in the site. Has characteristics. This feature allows the user to select a word pair that is truly associated with the web page as a search term.

  Furthermore, according to the present invention, when a search word candidate is presented by combining words, the candidate is determined by calculation in real time, so even if the search history is not accumulated, the hit rate is high. It is possible to reliably present a combination of search terms having a high value.

  Further, the above processing (for example, processing according to the flowcharts shown in FIGS. 7a to 7g and FIG. 8 above) is stored as a computer program in a storage medium such as a CD-R, ROM, DVD-ROM, MO, etc. By causing the computer to read (install or copy) the program stored in the storage medium, the computer can perform the above processing. Therefore, it is obvious that a computer program and a storage medium are also within the scope of the present invention.

It is a block diagram which shows an example of the whole structure of the search system in a website corresponding to embodiment of invention. It is a block diagram which shows an example of the hardware constitutions of the client 101 corresponding to embodiment of invention. It is a block diagram which shows an example of the hardware constitutions of the server 103 corresponding to embodiment of invention. It is a figure which shows an example of the word table corresponding to embodiment of invention. It is a figure which shows an example of the kana conversion table corresponding to embodiment of invention. It is a figure which shows an example of the Romaji conversion table corresponding to embodiment of invention. It is a state transition diagram showing an example of site search processing corresponding to an embodiment of the invention. It is a state transition diagram showing a processing example of the client 101 in the initial state corresponding to the embodiment of the invention. It is a state transition diagram showing a processing example of the client 101 in a search screen display waiting state corresponding to the embodiment of the invention. It is a state transition diagram showing a processing example of the client 101 in the search screen display state corresponding to the embodiment of the invention. It is a state transition diagram showing a processing example of the client 101 in a search word candidate reception waiting state corresponding to the embodiment of the invention. It is a state transition diagram showing a processing example of the client 101 in a search result reception waiting state corresponding to the embodiment of the invention. It is a state transition diagram which shows the example of a process of the server 103 in the request waiting state corresponding to embodiment of invention. It is a flowchart which shows an example of the search word candidate search process of FIG. It is a figure which shows an example of a search screen.

Claims (5)

In response to a search instruction from a client device that accesses a website published on the Internet, the server device performs a search in the website,
A first database for performing a full text search of information of web pages constituting the website;
A second database in which words contained in the web page are registered;
A receiving means for receiving an input word input for searching the website in the client device each time a new character is input;
Search means for searching the second database for a word that matches the received input word and determining a search word candidate;
Transmission means for transmitting the search word candidate determined by the search means to the client as a search word;
Search means for performing a full-text search using the first database based on the search terms received from the client device;
When the input word received by the receiving means includes a delimiter,
The search means includes
Searching the second database for a first word that matches the first input word immediately before the delimiter;
Searching for the first web page containing the searched first word using the first database;
Extracting a first neighboring word existing around the first word in the searched first web page;
The server apparatus, wherein the first word and the first neighboring word are determined as the search word candidates. The search means further includes:
The distance between the first word and the first neighboring word in the first web page, the appearance frequency of the first neighboring word in the first web page, the first word and A combination of the first word and the first surrounding word based on at least one of a selection frequency in which a combination with the first surrounding word is selected in the past and a ranking given to the combination. Calculate the first point given to
The server device according to claim 1, wherein the transmission unit transmits the first point together with the search word candidate to a client device. If the second input word is included immediately after the delimiter,
The search means further includes:
Selecting a second peripheral word that matches the second input word forward among the first peripheral words;
The server device according to claim 1 or 2, wherein the first word and the second neighboring word are determined as the search word candidates. The second database is
A word table for registering the word in association with a second point given based at least on the number of occurrences of the word on the website;
A conversion table for registering the hiragana notation of the word or the Roman alphabet notation,
The search means searches for the first word using the word table and the conversion table,
4. The server device according to claim 1, wherein the word is composed of a combination of at least one of kanji, hiragana, katakana, alphabet, symbol, number, and composite word. 5.   The computer program for functioning a computer as a server apparatus of any one of Claims 1 thru | or 4.