JPH1185737A - Device and method for managing dictionary and recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To efficiently process specified data at one part of unknown word data by performing prescribed processing to the unknown word data corresponding to specified managing data extracted from a dictionary. SOLUTION: A user can easily delete the unknown word data from the dictionary by specifying conditions for deleting unknown words. As a result, the state of registering a lot of unwanted unknown words on the dictionary and lowering the retrieval efficiency of the dictionary can be prevented. Further, the convenience of a Japanese word input device can be improved as well. Besides, since the conditions for deleting the unknown words can be specified based on the date of unknown word registration or the conditions of reference, the user can objectively judge whether each unknown word is required or not. Namely, only the unknown word with data, which show the reference conditions, showing the extremely small number of times can be deleted as well and only the unknown word showing the extremely old date of registration can be deleted as well. Further, by designating the both, only the unknown word showing the old date of registration and the low conditions of reference can be deleted as well.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a technique for managing a dictionary referred to in a Japanese input device. More specifically, the present invention relates to a technique for efficiently identifying a part of unknown word data registered in the dictionary. Processing technology.

[0002]

2. Description of the Related Art Conventionally, as a Japanese input device, various kana / kanji conversion devices for converting a kana character string input from a keyboard or the like into a desired kana / kanji mixed sentence, or a Japanese sentence taken by a scanner or the like. A so-called OCR device for recognizing characters from image information has been proposed. The kana-kanji conversion device searches for a kana-kanji mixed notation corresponding to the input kana character string by referring to a dictionary prepared in advance, and converts the kana character string into each notation. OC
In order to improve the conversion accuracy from image information to character data, that is, the literacy rate, the R device performs not only conversion on a character-by-character basis, but also an input Japanese character string on a word-by-word basis while referring to a dictionary. Some perform conversion to unusual characters.

Therefore, in order to input Japanese correctly with these Japanese input devices, it is important that abundant words are registered in the dictionary, but all words actually used are registered. It is very difficult to create a registered dictionary. This is because the terms used by the user of the kana-kanji conversion device vary depending on the contents input by the user, and various words are created as new words in daily life. Furthermore, it is almost impossible to register an address, a name, a product name, and the like in a dictionary. In order to solve these problems and ensure convenience for the user, many kana-kanji conversion devices prepare a dictionary in which only basic words are registered in advance, and for words that do not exist in the dictionary, , The user can additionally register a new word, that is, an unknown word. Further, a device for automatically detecting an unknown word and automatically registering it in a dictionary has been proposed (Japanese Patent Laid-Open No. 6-12453). When automatically registering an unknown word, a device that estimates all possible readings of the unknown word and registers the unknown word in several places in a dictionary has also been proposed.

On the other hand, there is a dictionary management apparatus for managing a dictionary in which unknown words can be additionally registered. The dictionary management device is a device for the user to perform processing such as deletion of unknown words additionally registered in the dictionary and manage the dictionary. Dictionary management is necessary because, when a large number of unknown words are additionally registered in the dictionary, the data to be referred to when searching the dictionary increases, thereby increasing the memory required for the dictionary and increasing the dictionary search efficiency. This is because it will decrease. In particular, in a device in which unknown words are automatically registered, unknown words that the user does not intend are registered in the dictionary, which may cause a reduction in search efficiency. In a conventional dictionary management device, a user selects and deletes registration data unnecessary for himself / herself while referring to a list of words registered in a dictionary.

[0005]

However, in the conventional dictionary management apparatus, little consideration has been given to the convenience in managing the dictionary. Taking the deletion of registration data as an example, as described above, the user needs to select and delete unnecessary registration data one by one from words registered in the dictionary. When the number of words is large, a lot of trouble is required for deletion. Also,
When a plurality of readings are given for one unknown word and are automatically registered in several places in the dictionary, it is difficult to find out and delete all of these data. That is, although there is a possibility that an erroneous reading is included in the plurality of readings, it has been difficult to delete data corresponding to the erroneous reading. Furthermore, when multiple readings are given and registered automatically, the dictionary cannot be easily managed because the user cannot know how many kinds of readings the unknown word is registered. . As described above, the difficulty in managing the dictionary has led to a reduction in the dictionary search efficiency, and has a possibility of reducing the convenience of the Japanese input device that refers to the dictionary. Note that such a problem has occurred not only when an unknown word that has been additionally registered is simply deleted but also when various processes for managing a dictionary are executed.

SUMMARY OF THE INVENTION The present invention has been made to solve at least a part of the above problems, and efficiently processes and manages a part of unknown word data in a dictionary in which unknown words can be additionally registered. The purpose is to provide a technology for performing.

[0007]

Means for Solving the Problems and Their Functions and Effects In order to solve at least a part of the problems described above, the present invention has the following configuration. A first dictionary management device according to the present invention includes:
A dictionary management device for managing a dictionary referred to in a Japanese input device for inputting a Japanese character string, wherein unknown word data that may include predetermined management data includes data on unknown words that do not exist in a prepared dictionary. A dictionary that can be additionally registered as, a condition for selecting at least a part of unknown words among the unknown words additionally registered in the dictionary, an unknown word search condition specifying unit that specifies based on the management data, The gist of the present invention is to include an unknown word extracting means for extracting unknown word data corresponding to the managed data from the dictionary, and an unknown word processing means for performing a predetermined process on the extracted unknown word data.

A first dictionary management method according to the present invention is configured so that data relating to an unknown word which does not exist in a prepared dictionary can be additionally registered as unknown word data which can include predetermined management data. A dictionary referred to in a Japanese input device for inputting a character string, a dictionary management method for managing by a computer, a condition for selecting at least some unknown words among unknown words additionally registered in the dictionary, The gist is to identify the unknown word data corresponding to the identified management data based on the management data, extract the unknown word data from the dictionary, and perform a predetermined process on the extracted unknown word data.

In the dictionary management apparatus and method, at least some of the unknown words added to the dictionary are selected based on the management data included in the unknown word data, and the selected unknown word data is added to the selected unknown word data. On the other hand, processing can be performed. At this time, if the management data is specified,
Since the process is performed after automatically extracting the unknown word data to be processed from the dictionary, it is not necessary for the user to select the unknown words one by one and perform the process.
Further, since all the unknown word data corresponding to the specified management data can be automatically extracted, the unknown word data can be completely processed, and the dictionary can be managed reliably.

The management data includes an index indicating an unknown word, data indicating the time when the unknown word was registered in the dictionary, data indicating the reference status of the unknown word after being registered in the dictionary, and the like. Various data can be considered. When the management data is registered in the dictionary, it is not necessary to be special data attached only to the unknown word. For example, it is assumed that the word class data of the word is used as the management data, and the word class “unknown word” is included in the word class data. May be included.

The above-mentioned predetermined process includes various processes performed on unknown word data. For example, the following unknown word data may be deleted, or a process of pseudo-deleting the unknown word data from the dictionary by attaching predetermined flag data to the corresponding unknown word data may be used. Further, a process of changing the management data attached to the unknown word data, such as resetting the reference status of the unknown word, may be used. further,
A process of deleting the flag from the unknown word data pseudo-deleted from the dictionary with the flag data attached thereto and restoring the unknown word data as valid unknown word data may be included.

[0012] Among these various data, the unknown word processing means is means for performing, as the predetermined processing, a process of deleting the extracted unknown word data from the dictionary. It is preferable that the data indicates the reference status of each unknown word data registered in the dictionary or the data indicates the time when each unknown word is registered in the dictionary.

If such data is used as management data, the user objectively determines whether or not each unknown word is unnecessary based on the reference status of the unknown word and the registered time, and deletes it. be able to. As a result, it is possible to delete only an unknown word whose data indicating the reference status indicates a very low number of times, or to delete only an unknown word whose registration time is very old. Further, both of the management data may be used. In this case, it is also possible to delete only unknown words whose registration time is old and whose reference status is low.

In the dictionary management apparatus described above, the dictionary includes a morphological analysis dictionary referred to in morphological analysis for separating Japanese character strings into phrases and a kana-kanji conversion dictionary referred to in kana-kanji conversion. The dictionary for morphological analysis is a dictionary in which the management data, the reading and kana-kanji mixed notation including unknown word data including the kana-kanji mixed notation are registered, and the kana-kanji conversion dictionary uses a reading composed of kana characters as an index. A dictionary in which kana-kanji mixed expressions are registered, wherein the unknown word extracting means extracts unknown word data corresponding to the specified management data from the morphological analysis dictionary; and Means for extracting unknown word data registered in the kana-kanji conversion dictionary based on the pronunciation and kana-kanji mixed notation included in the data. It may be what you get.

The dictionary management apparatus extracts unknown word data to be processed based on the management data from the morphological analysis dictionary. In the morphological analysis dictionary, the management data, the reading and the unknown word data including the kana-kanji mixed notation are registered in association with each other. Therefore, even when a plurality of readings are given to the unknown word to be processed, all the readings are given. Reading can be detected. Based on all the readings thus detected, the dictionary management device searches the kana-kanji conversion dictionary and extracts unknown word data to be processed. Therefore, according to such a dictionary management device, even when a plurality of readings are given to an unknown word and registered, and the unknown word data to be processed is scattered in several places in the kana-kanji conversion dictionary, it is efficiently used. Unknown words can be processed. If you try to retrieve unknown word data from the Kana-Kanji conversion dictionary without going through the morphological analysis dictionary described above, it will be necessary to search the data contained in the Kana-Kanji conversion dictionary one by one from the beginning to the end. is there.

In order to search for unknown word data registered in the kana-kanji conversion dictionary via the morphological analysis dictionary, the position of the unknown word data in the kana-kanji conversion dictionary is indicated in the morphological analysis dictionary instead of reading the unknown word. Special data may be provided. However, according to the above-described method of searching the kana-kanji conversion dictionary using the pronunciation registered in the morphological analysis dictionary as a medium, the morphological analysis dictionary is compared with a case where special data for searching the kana-kanji conversion dictionary is provided. This has the advantage that the memory capacity can be saved, and the management of both dictionaries is facilitated.

A second dictionary management device according to the present invention is a dictionary management device for managing a dictionary referred to in a Japanese input device for inputting a Japanese character string, and includes an unknown word that does not exist in a prepared dictionary. A dictionary capable of additionally registering data related to unknown words as unknown word data, reading detection means for detecting a reading input in the kana-kanji conversion by the Japanese input device for one unknown word registered in the dictionary, and the dictionary And unknown word extracting means for extracting unknown word data relating to the one unknown word, and performing a predetermined process on data corresponding to a reading different from the detected reading among the extracted unknown word data. The gist is to include an unknown word processing unit.

A second dictionary management method according to the present invention is configured such that data relating to an unknown word which does not exist in a prepared dictionary can be additionally registered as unknown word data. A dictionary referred to in the input device, a dictionary management method for managing by a computer, for one unknown word registered in the dictionary, to detect the reading input in the kana-kanji conversion by the Japanese input device, Extracting, from the dictionary, unknown word data related to the one unknown word, and performing a predetermined process on data corresponding to a reading different from the detected reading among the extracted unknown word data. And

In the dictionary management device and method, when an unknown word registered in the dictionary is referred to in the kana-kanji conversion by the Japanese input device, the data corresponding to the reading different from the reading input in the kana-kanji conversion is converted. Then, predetermined processing is performed. When an unknown word is automatically registered in the dictionary, a plurality of readings of the unknown word may be estimated and registered. In such a case, some of the estimated readings may be incorrect readings. According to the dictionary management device, by detecting the reading used in the kana-kanji conversion, without imposing a special burden on the user, detecting the unknown word data registered with such an incorrect reading, It can be processed efficiently.

The deletion of the unknown word data may be performed when the unknown word is referred to in the kana-kanji conversion. Alternatively, the reading referred to in the kana-kanji conversion may be separately saved. It may be performed after completing. Also, by asking the user to confirm when deleting data, or by referring to the management data included in the unknown word data by combining with the first dictionary management device, the detected reading and Only a part of data corresponding to different readings may be processed. By doing so, it becomes possible to appropriately manage data for unknown words that can be actually read.

[0021] In the above-mentioned second dictionary management device, the dictionary is a morphological analysis dictionary that is referred to in a morphological analysis in which a Japanese character string is divided into phrases by the Japanese input device, A kana-kanji conversion dictionary referred to in the kana-kanji conversion, and the morphological analysis dictionary is a dictionary registered in association with unknown word data including pronunciation and kana-kanji mixed notation, and the kana-kanji conversion dictionary is A dictionary in which kana-kanji mixed notation is registered as an index using readings composed of kana characters, and the unknown word extracting means is configured to register the unknown word registered in the morphological analysis dictionary based on the kana-kanji mixed notation of the unknown word. Means for extracting data, and the kana-kanji conversion dictionary based on the pronunciation and kana-kanji mixed notation included in the extracted unknown word data. The recording is unknown word data to be as including means for extracting.

As described in the first dictionary management device, if the unknown word data is to be extracted by the above method, a plurality of readings are given to the unknown word and registered, and the unknown word data is converted to kana-kanji conversion. Even if it is scattered in several places in the dictionary, unknown words can be processed efficiently. Further, since the search is performed using the reading registered in the morphological analysis dictionary as a medium, the memory capacity of the morphological analysis dictionary can be saved.

The present invention described above can be implemented using a computer. Therefore, the present invention
As described below, an embodiment as a recording medium in which programs for realizing various functions by a computer are recorded can be adopted.

The first recording medium of the present invention is configured so that data relating to unknown words that do not exist in a dictionary prepared in advance can be additionally registered as unknown word data that can include predetermined management data. A dictionary referred to when inputting a column, a computer-readable recording medium recording a program managed by the computer,
A function for specifying a condition for selecting at least a part of unknown words among the unknown words additionally registered in the dictionary, based on the management data, and unknown word data corresponding to the specified management data, And a function of performing a predetermined process on the extracted unknown word data by a computer.

In this recording medium, the predetermined process is a process of deleting the extracted unknown word data from the dictionary, and the management data is a reference to each unknown word data registered in the dictionary. It is desirable to use a recording medium that records a program that is data indicating a situation or data indicating a time when each unknown word is registered in the dictionary.

In this recording medium, among the data forming the dictionary, the management data, and the unknown word data including the kana-kanji mixed notation are registered as the unknown word extraction function. A function of operating data included in a morphological analysis dictionary referred to in a morphological analysis that separates word character strings into phrases, and extracting unknown word data corresponding to the specified management data from the morphological analysis dictionary. In the data forming the dictionary, the kana-kanji mixed notation is registered using the reading of kana characters as an index, and the data included in the kana-kanji conversion dictionary referred to in the kana-kanji conversion is operated. Function, based on the pronunciation and kana-kanji mixed notation included in the extracted unknown word data,
The recording medium may store a program having a function of extracting unknown word data registered in the kana-kanji conversion dictionary.

[0027] The second recording medium of the present invention is configured such that data relating to an unknown word that does not exist in a prepared dictionary can be additionally registered as unknown word data, and is referred to when a Japanese character string is input. Is a computer-readable recording medium on which a program managed by a computer is recorded, and reads, for one unknown word registered in the dictionary, a reading input in kana-kanji conversion by the Japanese input device. A function for detecting, from the dictionary, an unknown word extraction function for extracting unknown word data related to the one unknown word, among the extracted unknown word data,
A storage medium storing a program for realizing, by a computer, a function of performing predetermined processing on data corresponding to a reading different from the detected reading.

In this recording medium, as the unknown word extraction function, among the data forming the dictionary, unknown word data including readings and kana-kanji mixed notation are registered in association with each other, and Japanese character strings are registered. Extracts unknown word data registered in the morphological analysis dictionary based on the function of manipulating data contained in the morphological analysis dictionary referred to in the morphological analysis to be separated into phrases and the kana-kanji mixed notation of unknown words. In the data forming the dictionary, the kana-kanji mixed notation is registered using the reading of kana characters as an index, and the data included in the kana-kanji conversion dictionary referred to in the kana-kanji conversion is operated. The kana-kanji conversion dictionary based on the function and the kana-kanji mixed notation included in the extracted unknown word data. It may be a recording medium for recording a program and a function of extracting an unknown word data registered.

When the programs recorded on the recording media described above are executed by a computer to realize the respective functions, the unknown word registration device described above can be configured.

As a recording medium, a flexible disk, a CD-ROM, a magneto-optical disk, an IC card,
Various computer-readable media such as a ROM cartridge, a punched card, a printed matter on which a code such as a barcode is printed, an internal storage device (memory such as RAM and ROM) and an external storage device of the computer can be used. In addition, an aspect in which a computer program for causing a computer to realize the functions of each step or each unit of the above invention is supplied via a communication path, that is, the program is placed on a server or the like on a network, and a necessary program is provided via the communication path. May be downloaded to a computer and executed.

[0031]

DESCRIPTION OF THE PREFERRED EMBODIMENTS Embodiments of the present invention will be described below based on examples. FIG. 1 is a block diagram showing control logic of a Japanese input device including the dictionary management device of the present embodiment, and FIG. 2 is a block diagram showing hardware of the Japanese input device. First, a schematic configuration of hardware will be described with reference to FIG. 2 for convenience of description.

(1) Schematic Configuration of the Embodiment As shown in FIG.
20, ROM 22, RAM 24, hard disk 26,
The CD-ROM drives 27 are interconnected by a bus 38. The bus 38 has an input / output port 2
8 is also connected. A keyboard 30, a mouse 31, a CRT display 32, a printer 34, and a scanner 36 are connected to the input / output port 28 as input / output devices. The CD-ROM drive 27 can be a recording medium reading device corresponding to a recording medium in which a program is stored. For example, when a so-called floppy disk is used as a recording medium, a CD-R
Instead of the OM drive 27 or together with the CD-ROM drive 27, a floppy disk drive
May be connected. Also, the input / output port 28
Is connected to an external network NW via a public telephone line. Therefore,
The Japanese input device can download necessary programs and data to the hard disk 26 from the server SV also connected to the external network NW.

The CPU 20 is a well-known CPU.
M22 is a mask memory for storing basic software, etc.
The RAM 24 is a readable and writable memory constituting a main memory. The hard disk 26 stores a kana-kanji conversion program and other various programs loaded and executed in the RAM 24, and various conversion dictionaries referred to by the program. The kana-kanji conversion program may be stored in the ROM 22 or may be stored in the CD-ROM and stored in the CD-R.
The data may be read into the RAM 24 via the OM drive 27. In addition, the server SV sends the hard disk 26
May be downloaded.

With the hardware configured as described above,
Various functions such as text input, kana-kanji conversion, editing, display, printing, and dictionary management are realized. The text may be input from the keyboard 30 in the form of a character string, or may be input from the scanner 36 as image information. The sentence thus input is subjected to various processes described later by the CPU 20, stored in a predetermined area of the RAM 24, displayed on the screen of the CRT 26, and
4, etc. Various inputs for managing the dictionary are made from the keyboard 30 or the mouse 31.

Next, the function of each part when the Japanese input device of this embodiment is captured by functional blocks will be described with reference to FIG. Most of the functional blocks shown in FIG.
The PU 20 executes based on software.

The input unit 40 corresponds to the keyboard 30, the mouse 31, and the scanner 36 shown in FIG. 2, and inputs commands for inputting Japanese sentences and managing a dictionary. First, a case where a Japanese sentence is input will be described.

The text input from the input unit 40 is sent to the character receiving unit 48 under the control of the input / output control unit 46.
Here, when the keyboard 30 is operated, for example, the input / output control unit 46 performs a predetermined interrupt process on the CPU 20 to execute a character string input process.
When a text is input from the scanner 36, the driver of the scanner 36 is started.

The sentence thus input is subjected to morphological analysis by the morphological analyzer 50. What is morphological analysis?
For example, it refers to a process of analyzing a kana character string input as "Kuru-Hakoko / Hakobu" as "Kuru-Hako / Hakobu" by referring to the part of speech information of each word registered in the dictionary. . At this time, the morphological analysis unit 50 stores in the memory (ROM
22, the RAM 24, the hard disk 26) and the morphological analysis dictionary 62 and the kana-kanji conversion dictionary 64. The phrase candidate and the word candidate obtained during the course of the analysis are sent to the phrase candidate storage unit 54 and the word candidate storage unit 58, respectively, and stored in the RAM 24. When a kana character string has been input, the morphological analysis unit 50 performs kana-kanji conversion based on the morphological analysis result.

The candidates obtained during the morphological analysis in the kana-kanji conversion and stored in the phrase candidate storage unit 54 and the word candidate storage unit 58 are sent to the display unit 44 via the input / output control unit 46. Is displayed. Since these character strings may not be the desired character strings, the morphological analysis unit 50 performs processing such as displaying and selecting the next candidate in response to an instruction from the user. Although not shown, the results of these instructions and selections are stored as learning results.

On the other hand, after the morphological analysis unit 50 completes the morphological analysis of the sentence, the result is sent to the morphological output unit 52. The morpheme output unit 52 further outputs the result to the output unit 42 or the display unit 44 via the input / output control unit 46.
The output unit 42 corresponds to the printer 34 of FIG.
4 corresponds to the CRT display 32 of FIG. The output unit 42 also corresponds to a part that outputs the input Japanese character string to another application.

The result of the morphological analysis is also passed from the morphological output unit 52 to the unknown word extracting unit 56. Unknown word extraction unit 5
6 is a morphological analysis dictionary 62 based on the morphological analysis result.
An unknown word that does not exist in the kana-kanji conversion dictionary 64 is extracted. The unknown words thus extracted are stored in the unknown word registration unit 60.
Handed over to The unknown word registration unit 60 registers the unknown word in the morphological analysis dictionary 62 or the kana-kanji conversion dictionary 64 after performing a predetermined process such as the determination of the part of speech for each unknown word.

Next, a case where the dictionary is managed will be described. Commands for managing the dictionary, conditions for deleting unknown words (hereinafter referred to as deletion conditions), and the like are input from the input unit 40. The input deletion condition is sent to the deletion condition receiving unit 74 under the control of the input / output control unit 46, and further sent to the unknown word searching unit 76. The unknown word search unit 76 searches the morphological analysis dictionary 62 and the kana-kanji conversion dictionary 64 and extracts unknown word data corresponding to the input condition. This result is sent to unknown word deletion section 72. The unknown word deletion unit 72 accesses the morphological analysis dictionary 62 and the kana-kanji conversion dictionary 64 and deletes these data. Morphological analysis dictionary 62, kana-kanji conversion dictionary 64,
The unknown word deletion unit 72, the unknown word search unit 76, and the deletion condition receiving unit 74 constitute a dictionary management unit 70, and correspond to the dictionary management device in this specification.

(2) Unknown Word Registration Process Next, in order to clarify the format of unknown word data registered in the morphological analysis dictionary 62 and the kana-kanji conversion dictionary 64 managed by the dictionary management apparatus in this embodiment, An automatic registration process of an unknown word performed by the word input device will be described with reference to FIG. FIG. 3 is a flowchart showing the flow of the unknown word automatic registration process. This routine corresponds to the CP shown in FIG.
This is a process that is automatically performed by U20 while Japanese input is being performed. After Japanese input is finished,
It may be executed by inputting a predetermined command.

When the unknown word automatic registration routine is started,
The CPU 20 inputs a sentence (step S100),
A morphological analysis is performed (step S200). Regarding morphological analysis, various methods are known, such as a two-phrase longest-matching method in which two phrases are used as a basic unit and the longest phrase is obtained as a first candidate. In the embodiment, the minimum cost method is used. The minimum cost method is a method in which a cost is assigned to a word that can be a candidate for a word constituting a phrase and a combination of words, and a phrase whose score satisfies a predetermined condition is set as a first candidate.

Here, the morphological analysis method used in this embodiment will be described with reference to FIG. FIG. 4 is a flowchart showing the flow of the morphological analysis routine.
As described above, this is a routine for performing morphological analysis by a method called a minimum cost method. This processing is performed by the morphological analysis unit 50 based on the functional blocks in FIG. The morphological analysis unit 50 describes one processing function of the CPU 20 of FIG. 2 as a functional block.
It is the same meaning that U20 executes the morphological analysis routine.

As shown in FIG. 4, the CPU 20 first performs initialization (step S205) such as erasing temporarily stored data and initializing the analysis position to the first digit.
A process for obtaining an analysis position is performed (step S210). The analysis position is a position at which the input text is analyzed next. For example, assuming that a kana character string “Kuru wa Koko wa Kobu” is input, the first analysis position is the position of the first digit “ku”, and as the analysis proceeds, the analysis position becomes “ru”. "Ma" ... At this analysis position, the CPU 20 performs a process of searching the morphological analysis dictionary 62 and the kana-kanji conversion dictionary 64 stored in the hard disk 26 (step S215). In the previous example,
Search the dictionary for the word "ku".

After the dictionary is searched, the obtained word is checked for a connection with the previous word (step S220). If there is no connection between words, the word is invalidated. Further, the dictionary is searched. For example, in the previous example sentence (“Kuru-hanko-ko-kobu-bu”), “ha” of “ko-wo-ko-kobu” is searched for from the morphological analysis dictionary 62. "
Since it is determined that there is no possibility of combining with, it is treated as invalid data. Although not shown in the block diagram of FIG. 1, the connection between words is stored in a memory (RAM 24, ROM 22, hard disk 26) as a table indicating the possibility of connection depending on the part of speech. When the dictionary search and the connection check at one analysis position are completed, the analysis positions are sequentially advanced and the process is repeated.

For words that may be combined, see CP
U20 performs cost calculation, performs processing for obtaining the minimum total cost of the word (step S225), and performs processing for invalidating an inappropriate cost (step S23).
0). This is independent word =
2. This is a process in which the total cost up to the analysis position is calculated by defining it as having an attached word = 0 cost, and an inappropriate combination of costs that is larger than the combination of other words is invalidated. Explaining based on the example shown above, the "car" in the example sentence ("cars up to car") is "ku" +
It can be divided into various word combinations such as "ru" + "ma", "kuru" + "ma", "car". The cost is calculated by applying a word to each of these combinations.
If the word “bit” (independent word) + “style” (independent word) is applied to “ku” + “ru”, “style” has a cost of 4. On the other hand, if the word “kuru” (independent word) is applied to “kuru”, cost 2 is obtained. Since the minimum cost method employs a combination that minimizes the cost thus obtained, “coming” is employed in this case. If such analysis is continued, the cost of “car” (independent word) is “2” and the minimum cost is “car”.

Also, if we think about "until it comes,"
Because it is "car" (independent word) + "de" (attached word)
The cost of “de” is cost 2 corresponding to the total cost of “by car”. Similarly, considering the result of “to come” (independent word) + “to” (attached word), “to” also has a cost of 2.
The phrase candidates thus obtained and their costs are stored in the phrase candidate storage unit 54 (FIG. 1).

Next, a process of linking the word candidates to which the cost has been given in this way is performed (step S235). That is, by setting a pointer for a word for which the combination is valid, the combination is related. In the above example sentence, "to come" is described as "car / in" and "come /
Since the minimum total cost was calculated for "to", "coming" is linked to "to", and "car" is linked to "de".
The processing of such connection check, cost calculation, and linking is repeated until all words have been searched at one analysis position. When the dictionary search at the analysis position is completed, the analysis position is further advanced by one, the establishment of a new word is examined, and the connection check and cost calculation are repeated in the same manner.

When the analysis position has reached the position of the last kana character already input and the analysis has been completed for all words (step S240), the processing for searching for the path with the minimum cost is performed based on the above processing. (Step S24)
5). This is a process of searching for a combination of validated words that minimizes the sum of costs assigned to the words. In the example of "Have a car until coming", "car (2) / in (2) / box (4) / (4) / carry (6)"
Is the minimum cost with a total cost of 18.
The numbers in parentheses indicate the cost of each word.

At this time, although not the minimum cost, another segmentation candidate is also searched. For example, "car (2)
// (2) / is (2) / child (4) / (4) / carry (6) ". After creating the candidate for the segmentation in this way (step S25
0) This time, a process of creating a candidate inside each phrase is performed (step S255). That is, for example, candidates for "box" and "box" are prepared for "hakowo" within one segment break. These phrase candidates and word candidates are used when the user instructs to change the way of segmentation or instructs to display the next candidate.

In the above, the morphological analysis routine has been described by taking as an example the case where a kana character string is input. However, morphological analysis of a sentence containing katakana, kanji, alphabetic characters, numerals, and the like is the same processing. As is clear from the above description, in the morphological analysis, a dictionary search (step S21)
5) has an important role.

When the morphological analysis is completed, the CPU 20 proceeds to the next step and executes an unknown word extraction process (step S300). This is based on the result of the morphological analysis,
This is a process of extracting words that did not exist in the dictionary from the input text. In addition, the input character string may be determined in hiragana or katakana notation by performing a predetermined operation without depending on the morphological analysis routine, and such a word is also present in the dictionary. Those that do not are extracted as unknown words.

When the unknown word extraction processing is completed, an unknown word part of speech estimation processing for estimating the part of speech of the unknown word is performed (step S4).
00), a heading creation process for creating a reading to be registered in the dictionary is performed on the unknown word (step S500). On this occasion,
If the unknown word is a word consisting of kanji, the heading creation process generates a plurality of readings based on a combination of readings of the kanji. For example, in the case of an unknown word composed of two Chinese characters, four readings of “on-reading-on-reading”, “kun-reading-on-reading”, “on-reading-on-reading”, and “kun reading-on-reading” are generated. In the case of an unknown word consisting of three Chinese characters, the number of combinations will further increase. If a part of the unknown word exists in the dictionary, the pronunciation is generated using the pronunciation. For example, "birthday" is extracted as an unknown word,
If "Birthday" exists in the dictionary,
Only the "day" part is changed to various readings, and the readings "tanjo", "tanjohi", and "tanjo ni" are generated.

After the reading of the unknown word is generated by the above procedure, the unknown word is registered in the kana-kanji conversion dictionary 64 and the morphological analysis dictionary 62 (steps S600, S700).
Any of these registration orders may be performed first or may be performed simultaneously.

Here, the kana-kanji conversion dictionary 64 refers to a dictionary used in kana-kanji conversion, and uses an input kana character string as an index to mix kana-kanji characters (English characters,
(Including a mixture of numbers and symbols). When a plurality of readings of an unknown word are generated in the above-described headline creation process, the unknown word is registered in the kana-kanji conversion dictionary such that all the readings are used as an index so that the unknown word can be handled. Therefore, in the kana-kanji conversion dictionary 64, data corresponding to one unknown word can be registered without being associated with several places (see FIG. 7).

On the other hand, the morphological analysis dictionary 62 is a dictionary referred to in the morphological analysis described above (step S200). Therefore, unknown words are registered in the morphological analysis dictionary 62 in such a format that the part of speech of a word can be referred to using the character string as an index even if the input character string is mixed with kana-kanji characters (FIG. 7). reference). Also, unknown words are registered in the morphological analysis dictionary 62 in a form in which management data is attached. In the present embodiment, as management data, an index (“*” symbol in FIG. 7) indicating that the word is an unknown word, a registration date and a use frequency of the unknown word are registered together. The management data may be registered not only in the morphological analysis dictionary 62 but also in the kana-kanji conversion dictionary 64.

(3) Dictionary Management Processing Next, processing for dictionary management, which is a characteristic part of the present invention, will be described with reference to FIG. FIG. 5 is a flowchart showing the flow of the dictionary management routine of the present invention. This routine is executed by the CPU 20 shown in FIG. 2 by inputting a predetermined command separately from the Japanese input. However, the CPU 20 may execute the processing by performing a predetermined operation during Japanese input. In the following, an explanation will be given by taking an example of deleting unknown word data in the processing for managing the dictionary.

When the dictionary management routine is started, the CPU
20 displays a deletion deletion condition input screen for inputting a condition for specifying an unknown word to be deleted (step S80).
0). FIG. 6 shows an example of the deletion condition input screen. In the present embodiment, as a deletion condition corresponding to the management data, a condition as to whether or not batch deletion of unknown words, a condition based on a registration date of unknown words, and a condition based on reference frequency of unknown words can be designated. A condition based on the part of speech of an unknown word can be specified as a deletion condition corresponding to data. These conditions correspond to all input conditions (so-called AN
Only unknown word data (corresponding to the D condition) may be deleted, or unknown words that partially correspond to the input condition (corresponding to the so-called OR condition) may be deleted. In addition to the above-mentioned conditions, according to management data registered in the morphological analysis dictionary 62, deletion may be performed under various conditions, or unknown words may be individually specified and deleted. Is also good.

Next, the CPU 20 reads the deletion condition input by the keyboard 30 or the mouse 31 on the input screen of FIG. 6 (step S805 of FIG. 5), and retrieves the unknown word data corresponding to this condition from the morphological analysis dictionary 62. A search is performed (step S810). Whether an unknown word registered in the dictionary satisfies these deletion conditions is determined based on management data or the like registered in the morphological analysis dictionary 62. For example, in the case of batch deletion of unknown words, all data to which an index “*” symbol indicating an unknown word is attached is searched. For example, in the example illustrated in FIG. 7, the word “birthday” is determined to be an unknown word because “*” is attached in the morphological analysis dictionary 62.

Further, data indicating a registration date is also attached to the unknown word as management data. In the example shown in FIG. 7, the number “197070818” is registered as the registration date data. This means that “August 18, 1997” is the registration date. Therefore, when the registration date of the unknown word is specified as the deletion condition, the unknown word data corresponding to the deletion condition is extracted based on the value of the registration date data. Similarly, even when the reference frequency of an unknown word is specified, the corresponding unknown word data can be searched.

Next, the CPU 20 executes the morphological analysis dictionary 62
A search of the kana-kanji conversion dictionary 64 is performed based on the search result (step S815). The morphological analysis dictionary 62 includes:
Since all the readings are registered for the unknown word, the kana-kanji conversion dictionary 64 is searched using each reading registered here as an index. In the example shown in FIG. 7, when the unknown word “birthday” is searched as an unknown word to be deleted in the morphological analysis dictionary 62, the words “tanjobi”, “tanjohihi”, and “tanjo niichi” are found there. Yomi is registered. Therefore, next, data corresponding to these three types of readings is searched from the kana-kanji conversion dictionary 64. It should be noted that the kana-kanji conversion dictionary 64 should be used so that data with the same pronunciation but different kana-kanji mixed notation (for example, “comment” for the registered unknown word “kenkai”) is not extracted. In the search, not only the reading but also whether or not the kana-kanji mixed notation is the same as the notation registered in the morphological analysis dictionary 62 is determined at the same time.

After searching for the unknown word data corresponding to the deletion condition, the CPU 20 sets the morphological analysis dictionary 62
And these data are deleted from the kana-kanji conversion dictionary 64 (step S820). The unknown word data may be collectively deleted from the dictionary, or the user may be confirmed at the time of deletion.

According to such a dictionary management device, the user:
By specifying conditions for deleting unknown words, unknown word data can be easily deleted from the dictionary. As a result, since the dictionary can be easily managed, it is possible to prevent a state in which unnecessary unknown words are registered in the dictionary a lot and the search efficiency of the dictionary is reduced. This, in turn, improves the convenience of the Japanese input device. In addition, since the conditions for deleting unknown words can be specified based on the registration date of unknown words and the reference status of unknown words, the user can objectively determine whether each unknown word is unnecessary. can do. In other words, it is possible to delete only unknown words whose data indicating the reference status indicates a very low number of times, or to delete only unknown words whose registration time is very old. Also, by specifying both, the registration time is old,
In addition, it is also possible to delete only unknown words whose reference status is low.

On the other hand, according to the dictionary management apparatus, as a result of searching for unknown word data registered in the kana-kanji conversion dictionary 64 via the reading registered in the morphological analysis dictionary 62, a plurality of readings are given to the unknown word. Even if the unknown word data is scattered in several places in the kana-kanji conversion dictionary 64, the unknown word can be efficiently deleted. To search for unknown word data from the kana-kanji conversion dictionary 64 without going through the morphological analysis dictionary 62 described above, it is necessary to search the data contained in the kana-kanji conversion dictionary 64 one by one from the beginning to the end. This is because it occurs.

As described below, the dictionary management routine may be easily activated during the kana-kanji conversion in relation to the display in the kana-kanji conversion of the Japanese input device. For example, in the kana-kanji conversion, an unknown word is displayed in a form that can be clearly distinguished from other words (hereinafter, known words), and when the unknown word is displayed, by inputting a predetermined command, The kana-kanji conversion may be temporarily suspended to start the dictionary management routine.

For example, in the kana-kanji conversion, it is assumed that kana character strings of "Tango" and "Fukugo" are input as shown in the input screen (a) or the input screen (b) of FIG. At this time, referring to the dictionary data,
It can be seen that the word "Fukugo" is an unknown word and the word "Tango" is a known word depending on whether or not the symbol "*" is attached. Therefore, as shown in the output screens (a) and (b), the word "word" is displayed normally, and the unknown word "compound" is displayed in a frame. In addition, the display of the unknown word may be performed by changing the color, size, etc. of the character, or may be provided with an underline, hatching, or the like. Also,
Even when a plurality of conversion candidates are displayed, an unknown word may be displayed so as to be distinguishable from a known word. By inputting a predetermined command when the above-mentioned display is made, the dictionary management routine can be started. As the command, a predetermined function key of the keyboard 30 may be pressed, or a predetermined icon may be clicked with the mouse 31.

By displaying the unknown word in an identifiable manner, the user can easily notice the presence of the unnecessary unknown word if it is registered. Also, if the dictionary management function can be activated at the time of such display, the convenience of deleting unnecessary unknown word data and managing the dictionary is improved. In this case, by inputting a predetermined command, the unknown word displayed identifiable may be immediately deleted, or the dictionary management routine is started, and the deletion condition input screen shown in FIG. 6 is displayed. May be displayed.

In the above example, the deletion of unknown words has been described as an example of dictionary management processing. However, various processing can be performed by replacing step S820 in FIG. For example, skip flag data may be prepared in the management data, and the process of turning on the skip flag may be performed on the extracted unknown word data. When searching the dictionary in the kana-kanji conversion, if the word for which the skip flag is turned on is ignored, unknown word data can be deleted from the dictionary in a pseudo manner. According to such processing, by turning off the skip flag, it is also possible to restore the once deleted unknown word data to the dictionary as valid data again.

As another process, a process of initializing specific management data for the extracted unknown word data may be performed. For example, the data indicating the reference state of the unknown word (frequency data in FIG. 7) may be set to the value 0. This process is effective when the need for registered unknown word data is to be objectively evaluated again. This is because unknown word data having a low reference frequency is considered to be data of low necessity when a predetermined period has elapsed after initializing the reference state of unknown words. As described above, if the unknown word data can be appropriately extracted from the dictionary based on the management data, the dictionary can be managed very efficiently by preparing various processes according to the user's request. Will be able to do it.

(4) Dictionary Management Process as Second Embodiment Next, a dictionary management device as a second embodiment of the present invention will be described with an example in which unknown word data is deleted. The hardware configuration of the dictionary management device of the second embodiment is almost the same as that of the first embodiment described above (FIGS. 1 and 2). However, as shown by the broken line in FIG. 1, the morphological analysis unit 50 differs from the first embodiment in that the deletion condition of unknown words is also sent to the deletion condition receiving unit 74. In this embodiment, the reading of the unknown word is transmitted from the morphological analysis unit 50 to the deletion condition receiving unit 74 as the deletion condition.

FIG. 9 shows the flow of the dictionary management routine in the second embodiment. This routine is automatically executed by the CPU 20 even when the user does not input a predetermined command when an unknown word is referred to during Japanese input. However, it may be executed by the user inputting a predetermined command during or after the Japanese input.

When the dictionary management routine is started, the CPU
20 detects the reading of an unknown word used in kana-kanji conversion in Japanese input (step S900). If the dictionary management routine is executed in the middle of Japanese input, the hiragana character string input at the time of startup is read. Also,
If it is executed after Japanese input is finished, separate R
Data corresponding to reading an unknown word is read from the data stored in the AM 24 or the hard disk 26.

In general, in the kana-kanji conversion, a character string is inputted in a unit of a phrase or longer, so that
In order to detect the reading of an unknown word, it is necessary to analyze the input character string and write the phrase. The phrase segmentation is executed by the morphological analysis routine described above with reference to FIG.

Next, the CPU 20 searches the morphological analysis dictionary 62 for unknown word data referred to in the kana-kanji conversion (step S905). Since the morphological analysis dictionary 62 is not designed to search using the reading of an unknown word as an index (see FIG. 7), at this time, the dictionary is searched based on the referenced unknown word in kana-kanji notation. Therefore,
Either of steps S900 and S905 may be executed first.

Since the kana-kanji conversion dictionary refers to the kana-kanji conversion dictionary 64, the kana-kanji conversion dictionary 64 includes an index indicating the position of the data in the morphological analysis dictionary 62 corresponding to each data. The data may be registered in an attached form. By doing so, the processing in step S905 can be executed at high speed.

When the morphological analysis dictionary 62 is searched in this way, the CPU 20 reads the unknown word other than the reading detected in step S900 (hereinafter, referred to as “other reading”).
Is registered (step S91).
0). If no other reading is registered, it is determined that there is no unknown word data to be deleted from the dictionary,
20 temporarily ends the dictionary management routine. If another reading is registered, the process proceeds to the next step, and the CPU 2
0 executes a search of the kana-kanji conversion dictionary 64 (step S915). The readings other than the readings input in the kana-kanji conversion are presumed to be incorrect readings, so that these data are extracted and deleted.

Regarding the judgment in step S 910,
This will be described with a specific example of FIG. It is assumed that in the kana-kanji conversion, the reading "Tanjobi" is input and the unknown word data "Birthday" is referred to. The morphological analysis dictionary 62 includes:
In addition to the reading "Tanjobi", the readings "Tanjohi" and "Tanjohichi" are registered. Accordingly, since the other readings described above are registered, the CPU 20 searches the kana-kanji conversion dictionary to delete these data (step S91).
5).

The search of the kana-kanji conversion dictionary 64 is the same as the method described in step S815 of FIG. 5 in the first embodiment. However, in the second embodiment, only data corresponding to a reading other than the reading input in the kana-kanji conversion is detected. This is because the data to be deleted is data registered as an incorrect reading. When the unknown word data is searched in this way, the CPU 20 deletes the unknown word data (step S920). In particular,
In addition to deleting the retrieved data from the kana-kanji conversion dictionary 64, the morphological analysis dictionary 62 deletes reading data corresponding to another reading. The CPU 20 once terminates the dictionary management routine.

When an unknown word is given a plurality of readings and is automatically registered in a dictionary, it is presumed that some of the estimated readings are incorrect readings. According to the dictionary management device described above, unknown word data registered by such erroneous reading can be detected and efficiently deleted. In addition, since the dictionary management is automatically executed when the unknown word data is referred to in the kana-kanji conversion, the burden on the user in the dictionary management becomes very small. In addition, even if the user does not know how many readings the data is registered in, such as when multiple readings are automatically given and registered for unknown words, efficient dictionary management is possible. It can be carried out. Further, the second embodiment is different from the first embodiment in that the dictionary can be efficiently managed even for a dictionary in which unknown word data is registered without adding management data.

When deleting unknown word data, confirmation may be made with the user. Also, the management data registered in the morphological analysis dictionary 62 is referred to, and the reference frequency is 0.
Only data that satisfies a predetermined condition, such as the condition of times, may be deleted. This makes it possible to properly manage dictionaries for unknown words that have multiple readings (for example, unknown words such as “inset” can be read in two ways, “Ikan” and “Uchibame”). Will be able to do it. The second dictionary management device described above also displays unknown words in kana-kanji conversion so that they can be distinguished from other words, and then inputs a predetermined command,
The dictionary management routine may be activated.

In the above description, the deletion of unknown word data has been described as an example. However, various processes can be performed by replacing step S920 in FIG. 9 as in the first embodiment. That is, by turning on the skip flag described above, part of the read data is pseudo-deleted from the dictionary, or by turning off the flag, the read data is restored to the dictionary as valid. Or it may be done.

(5) Recording Medium as Dictionary Management Apparatus The dictionary management apparatus described above performs the various processes shown in FIGS. 5 and 9 on the computer shown in FIG.
0) is realized by execution, and therefore, the unknown word registration device can also be realized by a recording medium on which a program for realizing these functions is recorded. Such a recording medium may be a recording medium in which all of the unknown word automatic registration routines shown in FIGS. 5 and 9 are recorded, and the basic functions such as input / output functions of the above programs are described. Basic software (so-called BIOS or the like) separately provided in the computer may be used.

The recording medium may be a flexible disk, CD-ROM, magneto-optical disk, IC card,
Various computer-readable media such as a ROM cartridge, a punched card, a printed matter on which a code such as a barcode is printed, an internal storage device (memory such as RAM and ROM) and an external storage device of the computer can be used. In addition, an aspect in which a computer program that causes a computer to realize the functions of each step or each means of the above-described invention via a communication path, that is, as shown in FIG.
The program may be placed in a server SV or the like on the network NW, and the necessary program may be downloaded to a computer via a communication path and executed.

Although various embodiments of the present invention have been described above, the present invention is not limited to these embodiments, and various embodiments can be implemented without departing from the gist of the present invention.

[Brief description of the drawings]

FIG. 1 is a block diagram showing control logic of a Japanese input device including an unknown word registration device of the present embodiment.

FIG. 2 is a block diagram illustrating hardware of the Japanese input device;

FIG. 3 is a flowchart showing a flow of an unknown word automatic registration routine.

FIG. 4 is a flowchart showing a flow of a morphological analysis routine.

FIG. 5 is a flowchart showing the flow of a dictionary management routine.

FIG. 6 is an explanatory diagram showing an example of a deletion condition input screen.

FIG. 7 is an explanatory diagram showing a method of searching for unknown word data registered in a dictionary.

FIG. 8 is an explanatory diagram showing a display example of an unknown word.

FIG. 9 is a flowchart illustrating a flow of a dictionary management routine according to the second embodiment.

[Explanation of symbols]

 Reference Signs List 20 CPU 22 ROM 24 RAM 26 Hard disk 27 CD-ROM drive 28 Input / output port 30 Keyboard 31 Mouse 32 CRT Display 34 ... Printer 36 ... Scanner 38 ... Bus 39 ... Modem 40 ... Input unit 42 ... Output unit 44 ... Display unit 46 ... Input / output control unit 48 ... Character receiving unit 50: morphological analysis unit 52: morphological output unit 54: phrase candidate storage unit 56: unknown word extraction unit 58: word candidate storage unit 60: unknown word registration unit 62 morphological analysis dictionary 64 kana-kanji conversion dictionary 70 dictionary management unit 72 unknown word deletion unit 74 deletion condition receiving unit 76 unknown word search unit

Claims (12)

[Claims] 1. A dictionary management device for managing a dictionary referred to in a Japanese input device for inputting a Japanese character string, wherein data relating to unknown words that do not exist in a prepared dictionary is stored in a predetermined management data format. A dictionary that can be additionally registered as unknown word data that can be included, and an unknown word search condition specifying unit that specifies, based on the management data, a condition for selecting at least some unknown words among the unknown words additionally registered in the dictionary. And unknown word data corresponding to the specified management data,
A dictionary management device comprising: an unknown word extracting unit that extracts from the dictionary; and an unknown word processing unit that performs a predetermined process on the extracted unknown word data. 2. The dictionary management apparatus according to claim 1, wherein the unknown word processing unit is configured to perform, as the predetermined process, a process of deleting the extracted unknown word data from the dictionary. The dictionary management device, wherein the management data is data indicating a reference status of each unknown word data registered in the dictionary or data indicating a time when each unknown word is registered in the dictionary. 3. The dictionary management device according to claim 1, wherein the dictionary is a morphological analysis dictionary referred to in a morphological analysis that separates a Japanese character string into phrases, and a kana referred to in a kana-kanji conversion. A dictionary for kanji conversion, wherein the morphological analysis dictionary is a dictionary in which the management data, reading and unknown word data including kana-kanji mixed notation are associated and registered, and the kana-kanji conversion dictionary is composed of kana characters. A dictionary in which kana-kanji mixed notation is registered with the reading as an index, wherein the unknown word extraction means includes an unknown word data corresponding to the specified management data,
Means for extracting from the morphological analysis dictionary, and means for extracting unknown word data registered in the kana-kanji conversion dictionary based on the pronunciation and kana-kanji mixed notation included in the extracted unknown word data Dictionary management device. 4. A dictionary management device for managing a dictionary referred to in a Japanese input device for inputting a Japanese character string, wherein data relating to an unknown word that does not exist in a previously prepared dictionary is additionally registered as unknown word data. A possible dictionary, reading detection means for detecting a reading input in the kana-kanji conversion by the Japanese input device for one unknown word registered in the dictionary, and an unknown word related to the one unknown word from the dictionary. Dictionary management, comprising: unknown word extraction means for extracting word data; and unknown word processing means for performing a predetermined process on data corresponding to a reading different from the detected reading among the extracted unknown word data. apparatus. 5. The dictionary management device according to claim 4, wherein the dictionary is a morphological analysis dictionary that is referred to in a morphological analysis in which a Japanese character string is segmented into phrases by the Japanese input device. A kana-kanji conversion dictionary referred to in kana-kanji conversion by the word input device, wherein the morphological analysis dictionary is a dictionary registered in association with unknown word data including readings and kana-kanji mixed notations, and the kana-kanji The conversion dictionary is a dictionary in which kana-kanji mixed notation is registered by using readings composed of kana characters as an index, and the unknown word extracting means is registered in the morphological analysis dictionary based on the kana-kanji mixed notation of unknown words. Means for extracting the extracted unknown word data, and the kana kanji based on the pronunciation and kana kanji mixed notation included in the extracted unknown word data. Means for extracting unknown word data registered in the conversion dictionary. 6. A Japanese input device for inputting a Japanese character string, wherein data relating to an unknown word that does not exist in a prepared dictionary can be additionally registered as unknown word data that can include predetermined management data. A dictionary management method for managing by computer, wherein a condition for selecting at least some unknown words among the unknown words additionally registered in the dictionary is specified based on the management data, Unknown word data corresponding to the specified management data,
A dictionary management method that extracts from the dictionary and performs a predetermined process on the extracted unknown word data. 7. A dictionary which is configured so that data relating to an unknown word which does not exist in a previously prepared dictionary can be additionally registered as unknown word data, and which is referred to in a Japanese input device for inputting a Japanese character string, A dictionary management method for managing by a computer, wherein, for one unknown word registered in the dictionary, a pronunciation input in kana-kanji conversion by the Japanese input device is detected, and the one unknown word is detected from the dictionary. A dictionary management method for extracting unknown word data related to the data and performing a predetermined process on data corresponding to a reading different from the detected reading among the extracted unknown word data. 8. Data relating to an unknown word that does not exist in a dictionary prepared in advance can be additionally registered as unknown word data that can include predetermined management data, and is referred to when a Japanese character string is input. A dictionary which is a computer-readable recording medium storing a program managed by a computer, wherein a condition for selecting at least a part of unknown words among unknown words additionally registered in the dictionary is determined based on the management data. And unknown word data corresponding to the specified management data,
A recording medium storing a program for realizing, by a computer, an unknown word extraction function for extracting from the dictionary and a function of performing predetermined processing on the extracted unknown word data. 9. The recording medium according to claim 8, wherein the predetermined process is a process of deleting the extracted unknown word data from the dictionary, and wherein the management data is stored in the dictionary. A recording medium on which a program which is data indicating a reference state of unknown word data or data indicating a time when each unknown word is registered in the dictionary is recorded. 10. The recording medium according to claim 8, wherein, as the unknown word extraction function, among the data forming the dictionary, the management data, the unknown word data including the pronunciation and kana-kanji mixed notation are associated with each other. A function of operating data included in a morphological analysis dictionary that is registered and referred to in a morphological analysis that separates Japanese character strings into phrases, and unknown word data corresponding to the specified management data,
A function for extracting from the morphological analysis dictionary, and a kana-kanji mixed notation registered with a kana-kanji mixed notation as an index using a reading composed of kana characters, among data forming the dictionary, for kana-kanji conversion. A function of operating data included in the dictionary, and a function of extracting unknown word data registered in the kana-kanji conversion dictionary based on the pronunciation and kana-kanji mixed notation included in the extracted unknown word data. Recording medium on which a program is stored. 11. A computer which manages a dictionary referred to when inputting a Japanese character string, wherein data relating to an unknown word which does not exist in a previously prepared dictionary can be additionally registered as unknown word data. A computer-readable recording medium on which a program to be recorded is recorded, wherein, for one unknown word registered in the dictionary, a function of detecting a reading input in kana-kanji conversion by the Japanese input device; and An unknown word extraction function of extracting unknown word data related to the one unknown word; and a function of performing a predetermined process on data corresponding to a reading different from the detected reading, among the extracted unknown word data. Recording medium for recording a program for realizing the program by a computer. 12. The recording medium according to claim 11, wherein, as the unknown word extraction function, unknown word data including readings and kana-kanji mixed notation among data forming the dictionary is registered in association with each other. A function to manipulate data included in a morphological analysis dictionary referred to in a morphological analysis that separates a Japanese character string into phrases, and a function that is registered in the morphological analysis dictionary based on a kana-kanji mixed notation of an unknown word. A function for extracting unknown word data, and a kana-kanji conversion dictionary registered as a kana-kanji conversion notation, in which kana-kanji mixed notation is registered as an index with readings of kana characters among data forming the dictionary. Based on the function of operating the included data, and the pronunciation and kana-kanji mixed notation included in the extracted unknown word data, Recording medium for recording a program and a function to extract the unknown word data that has been registered in the name kanji conversion dictionary.