JP4238642B2 - Word registration device, word registration method, and word registration program - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention stores a word character string and its reading in association with each other, and registers a non-registered word character string and its reading in a word dictionary used for conversion from reading to word character string (notation) In addition, the present invention relates to a word registration method and a word registration program, such as a cellular phone, a portable electronic mail terminal, a PDA, and a personal computer.
[0002]
[Prior art]
For devices such as mobile phones, mobile e-mail terminals, PDAs, and personal computers where document input is performed, reading (Kana, Roman characters, etc.) input from the keyboard is converted into word character strings (Kanji, Katakana, etc.). Conversion programs (FEP, IME) are installed.
[0003]
This type of conversion program retrieves a character string corresponding to a reading from the word dictionary in accordance with a conversion operation after reading input, and displays this in an unconfirmed state. If the converted character string is correct, the input of the converted character string is confirmed according to the confirming operation. If the converted character string is incorrect, the conversion is performed according to the conversion candidate switching operation, the phrase delimiter changing operation, etc. Correct the string.
[0004]
Usually, the conversion program as described above has a word registration function in order to improve the conversion accuracy. The word registration function is a function for additionally registering unregistered words in the word dictionary. At this time, a word character string and its reading are input. Therefore, the user can perform highly accurate character string conversion by additionally registering unregistered words sequentially or collectively in the word dictionary.
[0005]
By the way, in recent years, with the spread of e-mail and the like, cases of inputting a document while referring to a document created by another user are increasing. In such a case, a part of the document to be referred to may be quoted. At this time, the user inputs the reading of the quoted part and converts it into a character string such as kanji.
However, if the above quoted part includes a word character string that is not normally used by the user, and the word character string is not registered in the dictionary, correct character string conversion is not performed, and the efficiency of sentence input decreases. There is a problem of doing.
[0006]
Thus, an unregistered word character string is automatically extracted from a sentence that has already been converted to a character string (see, for example, Patent Documents 1 and 2), and an unregistered word character string included in a document file is converted into a conversion candidate. Have been proposed (for example, see Patent Document 3), and those that automatically register unregistered word character strings and their readings in a word dictionary (for example, see Patent Document 4). .
[0007]
[Patent Document 1]
Japanese Patent Laid-Open No. 7-21170 (page 3, FIG. 2)
[Patent Document 2]
Japanese Patent Laid-Open No. 2002-229999 (page 9, FIG. 4)
[Patent Document 3]
Japanese Patent Application Laid-Open No. 10-334090 (page 4, FIG. 7)
[Patent Document 4]
Japanese Patent Laid-Open No. 11-85761 (page 9, FIG. 5)
[0008]
[Problems to be solved by the invention]
However, in Patent Documents 1 and 2, an unregistered word character string is automatically extracted from a sentence that has already been converted into a character string, but when registering the extracted word character string in a word dictionary, , The user needs to input the reading.
Note that Patent Documents 1 and 2 do not disclose a specific method for automatically extracting a word character string from a sentence.
[0009]
Also, in Patent Document 3, an unregistered word character string and its reading are extracted at the time of inputting a document, and the information is added to the document file so that another user can open the document file. When editing / modifying, unregistered word character strings are also displayed as conversion candidates based on the additional information. In other words, the technique disclosed in Patent Document 3 is based on the premise that special information is added to a document file, and there is a problem that the effect is not exhibited unless the user uses the same conversion program.
[0010]
Moreover, what is shown in Patent Document 4 is for automatically registering an unregistered word character string included in a character string input by a user together with its reading in a word dictionary. It does not extract word character strings from sentences such as e-mails converted into strings.
In addition, in the technique disclosed in Patent Document 4, since a character string is extracted using a complicated morphological analysis routine, if a converted document such as an e-mail is targeted, the processing burden of the apparatus There is a problem that increases. In particular, a small device such as a cellular phone with limited hardware resources is likely to be difficult to implement due to insufficient processing capacity.
[0011]
The present invention has been considered in view of the above circumstances, and is intended for sentences already converted into character strings, extracts unregistered word character strings, generates readings thereof, and extracts extracted word character strings and By registering the generated readings in the word dictionary, the word registration work by the user can be reduced, and by restricting the hardware resources by simplifying the word string extraction process and the reading generation process as much as possible An object of the present invention is to provide a word registration device, a word registration method, and a word registration program that can be implemented even in a small device such as a mobile phone.
[0012]
[Means for Solving the Problems]
In order to achieve the above object, the word registration apparatus of the present invention stores a word character string and its reading in association with each other and stores an unregistered word character string with respect to a word dictionary used for conversion from reading to word character string. And a word registration device for registering readings thereof, which target sentences that have already been converted into character strings and extract word character strings to be registered in the word dictionary using a word character string extraction keyword A character string extracting unit that decomposes the extracted word character string into character units, searches the word dictionary for readings of each decomposed character, combines the searched plural readings, and extracts the word characters A reading generation means for generating a reading of a sequence and a word registration means for associating the extracted word character string and the generated reading with each other and registering them in the word dictionary are provided.
[0013]
If the word registration device is configured in this way, a sentence that has already been converted into a character string is targeted, an unregistered word character string is extracted, its reading is generated, and the extracted word character string and the generated reading are read. It becomes possible to register in the word dictionary. Thereby, the word registration work by the user can be reduced, and the efficiency of sentence input can be improved when inputting a document while referring to a sentence that has already been converted into a character string.
[0014]
Also, by extracting a word character string using a word character string extraction keyword, the word character string extraction process can be simplified as compared with the case of using a complicated morpheme analysis routine.
Also, the reading generation processing can be simplified by searching for the reading of the extracted word character string in character units and combining the searched plural readings to generate the reading of the word character string.
As a result, not only can the processing load of the apparatus be reduced, and a quick word registration process can be performed, but also a small apparatus such as a mobile phone with limited hardware resources can be implemented.
[0015]
In the word registration device of the present invention, the word character string extraction unit extracts a word character string and / or a descriptive symbol sandwiched between the word extraction keywords and a word character string sandwiched between the word extraction keywords. Of the extracted word character strings, those not registered in the word dictionary are registered.
If the word registration device is configured in this way, the extraction process of the word character string is further simplified, so that not only the processing load can be reduced, but also the processing speed can be further improved.
[0016]
In the word registration device of the present invention, the word character string extraction unit extracts a word character string composed only of kanji and / or a word character string composed only of katakana.
If the word registration device is configured in this way, the extraction process of the word character string is further simplified, so that not only the processing load can be reduced, but also the processing speed can be further improved.
[0017]
In the word registration device of the present invention, the word dictionary stores word character string readings in units of word character string characters.
If the word registration device is configured in this way, the dictionary data for converting from reading to word character string and the dictionary data for generating the reading of the extracted word character string are combined to reduce the word dictionary capacity. Can do. This further facilitates implementation on a cellular phone or the like that has limited hardware resources.
[0018]
In the word registration device of the present invention, the word dictionary includes an original dictionary and a user dictionary used for conversion from reading to a word character string, and the reading generation unit reads the word character string extracted. Is performed using the original dictionary, and the word registration means registers words in the user dictionary.
If the word registration device is configured in this way, reading of the extracted word character string can be generated with high accuracy without being affected by the dictionary data registered in the user dictionary.
[0019]
In the word registration device of the present invention, the word registration unit displays the extracted word character string and the generated reading, and requests correction and / or registration confirmation.
If the word registration device is configured in this way, it is possible not only to prevent erroneously extracted word character strings or erroneously generated readings from being registered in the word dictionary, but also to register word words and their Users can recognize readings and perform efficient document input.
[0020]
Moreover, the word registration device of the present invention is configured to be built in a small communication terminal device.
Since the word character string extraction process and the reading generation process are simplified, the word registration device of the present invention can be implemented even in a small communication terminal device such as a mobile phone with limited hardware resources. Become.
[0021]
In order to achieve the above object, the word registration method of the present invention stores a word character string and its reading in association with each other, and stores unregistered words in a word dictionary used for conversion from reading to word character string. A word registration method for registering a character string and a reading thereof, for a sentence that has already been converted to a character string, and extracting a word character string to be registered in the word dictionary using a word character string extraction keyword Then, the extracted word character string is decomposed into character units, readings of each decomposed character are searched from the word dictionary, and a plurality of searched readings are combined to read the extracted word character string. The generated and extracted word character string and the generated reading are associated with each other and registered in the word dictionary.
[0022]
Further, the word registration method of the present invention is implemented in word registration in a small communication terminal device.
In this way, the word character string extraction process and the reading generation process can be easily performed, and thus can be performed even in a small communication terminal device such as a cellular phone in which hardware resources are limited. become.
[0023]
If such a word registration method is used, a sentence that has already been converted to a character string is extracted, an unregistered word character string is extracted, its reading is generated, and the extracted word character string and the generated reading are generated. Is registered in the word dictionary, so that the word registration work by the user can be reduced and efficient document input can be performed.
In addition, by simplifying the word character string extraction process and the reading generation process as much as possible, it is possible to implement even a small device such as a mobile phone with limited hardware resources.
[0024]
In order to achieve the above object, the word registration program of the present invention stores a word character string and its reading in association with each other and stores unregistered words in a word dictionary used for conversion from reading to word character string. A word registration program for registering a character string and its reading, which is a word registration device that targets a sentence that has already been converted to a character string, and is a registration target of the word dictionary using a word character string extraction keyword Extracting a word character string, decomposing the extracted word character string into character units, searching for the reading of each decomposed character from the word dictionary, combining the searched plural readings, and extracting the word A reading of a character string is generated, and the extracted word character string and the generated reading are associated with each other and registered in the word dictionary.
[0025]
If the word registration program is configured in this way, a sentence already converted to a character string is targeted, an unregistered word character string is extracted, its reading is generated, and the extracted word character string and the generated reading are read. Since it is registered in the word dictionary, the word registration work by the user can be reduced.
[0026]
The word registration program of the present invention is configured to execute word registration in a small communication terminal device.
In this way, word registration can be easily realized even in a small communication terminal device such as a mobile phone in which hardware resources are limited.
[0027]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
[0028]
[First embodiment]
First, a first embodiment of the present invention will be described with reference to FIGS.
FIG. 1 is a block diagram showing a hardware configuration of a document input device (word registration device) according to the first embodiment of the present invention.
[0029]
The document input device 1 shown in this figure is functionally provided with the word registration device of the present invention, and is constituted by a device having a document input function, such as a mobile phone, a portable electronic mail terminal, a PDA, a personal computer, for example. Has been.
For example, when the document input device 1 is a personal computer, the hardware includes an input unit 2 such as a keyboard, a display unit 3 such as a liquid crystal display, a communication unit 4 such as a LAN, and a storage unit 5 such as a hard disk. And a control unit 6 such as a CPU.
[0030]
FIG. 2 is a block diagram showing a functional configuration of the document input device (word registration device) according to the first embodiment of the present invention.
As shown in this figure, the document input device 1 functionally configures a document database 10, a Japanese language conversion device 20, a word registration device 30, and the like by a program stored in the storage unit 5.
The document database 10 stores document data created by users and document data created by other users. An example of document data created by another user is a received mail.
[0031]
The Japanese conversion device 20 includes a conversion engine 22 that converts a reading into a word character string using a word dictionary 21 that stores the word character string and its reading in association with each other. More specifically, the conversion engine 22 searches the word dictionary 21 for a character string corresponding to reading in accordance with the conversion operation after reading input in the input unit 2, and displays this on the display unit 3 in an unconfirmed state. To do. If the converted character string is correct, the input of the converted character string is confirmed according to the confirming operation in the input unit 2, while if the converted character string is not correct, the conversion candidate switching operation and phrase break in the input unit 2 are performed. The conversion character string is corrected according to the change operation.
[0032]
The word dictionary 21 includes an original dictionary 21a and a user dictionary 21b. The original dictionary 21a is a word dictionary that is provided as a standard in the document input device 1, and the user dictionary 21b is a word dictionary that allows additional registration of words by the user. In the original dictionary 21a of this embodiment, in order to perform a reading search of an extracted word character string, which will be described later, word character string readings are stored in units of word character string characters. For example, the word dictionary data “word character string: computer, reading: Keisanki” is internally configured as “word character string: total / calculation / machine, reading: Kei / san / ki”. It can also be used as single-kanji dictionary data such as “character: total, reading: kei”, “character: arithmetic, reading: san”, “character: machine, reading: ki”.
[0033]
The word registration device 30 includes a word character string extraction unit 31, a reading generation unit 32, and a word registration unit 33.
The word character string extraction means 31 targets sentence data in the document database 10 that has already been converted to a character string, and uses a word character string extraction keyword described later as a registration target of the word dictionary 21 (user dictionary 21b). This is a functional component that automatically extracts a word character string.
The word character string extraction means 31 of this embodiment has a function of presetting document data conditions to be extracted and word character string extraction conditions.
[0034]
The reading generation means 32 decomposes the extracted word character string into character units, searches the word dictionary 21 (original dictionary 21a) for the readings of the decomposed characters, and combines the extracted readings to extract them. It is a functional component that generates a reading of a word string.
The word registration unit 33 is a functional component that registers the extracted word character string and the generated reading in association with each other in the word dictionary 21 (user dictionary 21b).
The word registering means 33 of the present embodiment has a function of displaying the extracted word character string and the generated reading on the display unit 3 and requesting the user to make corrections or confirm registration.
[0035]
Next, the operation of the word registration device 30 in the present embodiment will be described with reference to FIGS.
FIG. 3 is a flowchart showing the operation of the word registration device according to the first embodiment of the present invention, and FIG. 4 is a word character string extraction used by the word registration device according to the first embodiment of the present invention to extract a word character string. It is explanatory drawing which shows the keyword for use.
[0036]
As shown in FIG. 3, first, the user inputs word character string extraction conditions from the input unit 2 (S101). The word character string extraction means 31 performs condition settings such as selection of document data to be extracted from the document database 10 and the number of word extractions according to the word character string extraction conditions input from the input unit 2, and according to the settings. Then, a word character string is extracted from the document data to be extracted (S102).
This word character string extraction processing is performed using a word character string extraction keyword as shown in FIG. For example, the word character string extraction keywords include case particles, case particle equivalents, ramification particles, collection particles, case particles + collection particles, collection particles + case particles, connection particles, determination particles, and the like.
[0037]
Specifically, it follows the word string extraction keyword, followed by the character string up to the word string extraction keyword and the descriptive symbols (punctuation marks, question marks, exclamation marks, etc.), and the word string extraction keyword. The character string up to is extracted.
Moreover, the word character string to be extracted is limited to a word character string composed only of kanji and a word character string composed only of katakana, and word characters to be registered in the word dictionary 21 (user dictionary 21b). The columns are word character strings that are not registered in the word dictionary 21 (original dictionary 21a and user dictionary 21b).
[0038]
Next, a reading of the extracted word character string is generated (S103). If the extracted word character string is katakana, the corresponding reading is used. If the extracted word character string is kanji, a word character string reading is generated using the word dictionary 21 (original dictionary 21a). .
As described above, in the original dictionary 21a, the reading is divided in character units of the word character string. Therefore, the extracted word character string is decomposed into character units, and the reading of each character is searched in the original dictionary 21a. Then, a plurality of retrieved readings are combined to obtain an extracted word character string reading. When there are a plurality of reading candidates, the reading with the largest number of hits is adopted.
[0039]
Next, the extracted word character string and the generated reading are displayed on the display unit 3 (S104), and a correction or registration confirmation is requested from the user (S105). Here, when correction is not necessary, the extracted word character string and the generated reading are registered in the word dictionary 21 (user dictionary 21b) (S106), and when correction is necessary, the user manually enters the word character string or After correcting the reading (S107), the word character string and the reading are registered in the word dictionary 21 (user dictionary 21b).
Thereafter, if the above reading registered in the word dictionary 21 (user dictionary 21b) is input, it can be converted into the above word character string (S108).
[0040]
Next, a specific operation example of the word registration device according to the first embodiment of the present invention will be described.
For example, in the document database 10, there is an incoming mail including a character string “Need to have a state system in the future.” In the original dictionary 21a, “word character string: road / road, reading: how / ro”, “word Consider a case where there is dictionary data such as “character string: book / state, reading: hon / shu” and “word string: control / degree, reading: sei / do”. First, the user selects a corresponding received mail (S101), and a word character string extraction process is performed on the received mail (S102). In this word string extraction step, for the sentence “In the future, a state system will be required.” Included in the received email, it is the string that follows the punctuation and before the word string extraction keyword “ga”. In addition, “Doshu system” is extracted as a character string whose character type is composed of only kanji.
[0041]
Next, a reading of the extracted character string “Doshu system” is generated (S103). In this reading generation processing step, the character string “state” is disassembled into character units, and the readings of the characters “way”, “state”, “system” are searched in the original dictionary 21a. In this example, “character: road, reading: how”, “character: state, reading: Shu”, “character: system, reading: sei” are searched, and these are combined to obtain dictionary registration candidate data “character string: "Doshu system, reading: how to say".
[0042]
Next, the extracted word character string and its reading are displayed to the user (S104), and the presence or absence of correction is confirmed (S105). In this case, since no correction is necessary, the dictionary registration candidate data “character string: state system, reading: doseisei” is registered in the user dictionary 21b in accordance with the user's confirmation operation (S106).
By executing the above steps, the next time the document is input, the conversion engine 22 adds a new registered word “word state system” to the user dictionary 21b for the input reading “Doshusei”. Can be retrieved and displayed as conversion candidates (S108).
[0043]
According to the present embodiment configured as described above, a sentence already converted to a character string is targeted, an unregistered word character string is extracted, its reading is generated, and the extracted word character string and generation It becomes possible to register the readings made in the word dictionary 21. Thereby, the word registration work by the user can be reduced, and the efficiency of sentence input can be improved when inputting a document while referring to a sentence that has already been converted into a character string.
[0044]
Also, by extracting a word character string using a word character string extraction keyword, the word character string extraction process can be simplified as compared with the case of using a complicated morpheme analysis routine. Also, the reading generation processing can be simplified by searching for the reading of the extracted word character string in character units and combining the searched plural readings to generate the reading of the word character string. As a result, not only can the processing load of the apparatus be reduced, and a quick word registration process can be performed, but also a small apparatus such as a mobile phone with limited hardware resources can be implemented.
[0045]
In addition, in the character string extraction process, a word character string consisting of only kanji or katakana following the word extraction keyword or description symbol and the word character string up to the word extraction keyword is extracted. Of the extracted word character strings, those not registered in the user dictionary 21b are to be registered, so that the word character string extraction process can be further simplified and the processing load can be reduced. Further improvement can be achieved.
[0046]
Further, the word dictionary 21 (original dictionary 21a) stores the reading of the word character string by dividing the word character string into character units, so that the dictionary data for converting the reading into the word character string and the extracted word characters are stored. The dictionary data for generating column readings can also be used in common, and the word dictionary capacity can be reduced. This further facilitates implementation on a cellular phone or the like that has limited hardware resources.
[0047]
The word dictionary 21 includes an original dictionary 21a and a user dictionary 21b that are used for conversion from reading to word character strings. The word dictionary 21 is used to perform a reading search of the extracted word character strings, and word registration is performed. Since it is performed on the user dictionary 21b, reading of the extracted word character string can be generated with high accuracy without being affected by the dictionary data registered in the user dictionary 21b.
[0048]
In addition, since the extracted word character string and the generated reading are displayed and the correction or registration confirmation is requested, the word character string extracted by mistake or the erroneously generated reading is displayed in the word dictionary 21 (user dictionary 21b). In addition, the user can recognize the word character string to be registered and the reading thereof, and can efficiently input a document.
[0049]
[Second Embodiment]
Next, a second embodiment of the present invention will be described with reference to FIG.
FIG. 5 is a flowchart showing the operation of the word registration device according to the second embodiment of the present invention.
In the second embodiment shown in this figure, the user directly selects the document data to be extracted from the document database 10, and the word dictionary 21 (user) without confirming the extracted word character string and its reading to the user. The point of registration in the dictionary 21b) is different from the above embodiment.
[0050]
As shown in FIG. 5, in the second embodiment, first, the user selects document data in the document database 10 (S201). Thereafter, the word character string extraction processing (S202), the extracted word character string reading generation processing (S203), and the extracted word character string and generated reading dictionary registration processing (S204) are automatically performed. Executed. After that, if the above reading registered in the word dictionary 21 (user dictionary 21b) is input, it can be converted into the above word character string (S205).
[0051]
In this embodiment, no correction request or registration confirmation is made to the user. Therefore, in the word character string extraction process and the word character string reading generation process, the extraction condition and the reading generation condition are stricter, and the dictionary has a high accuracy rate. It is preferable to extract and generate only data. For example, the word character string extraction processing uses only the word character string extraction keywords shown in FIG. 4 with a high accuracy rate. In the reading generation processing, the word character string reading search is performed in character units. By adding a condition such as using only the number of hits of the same reading candidate exceeding a predetermined threshold, a candidate with high accuracy is selected.
[0052]
Next, a specific operation of the second embodiment will be described.
For example, the document database 10 includes a received mail including a sentence “Tomorrow will discuss the state system”, and the original dictionary 21 a includes “character string: road / road, reading: how / ro”, "Character string: Ayumu / Road, Reading: Ho / Do", "Character string: Book / State, Reading: Hon / Shyu", "String: Kyu / State, Reading: Kyu / Shyu", "String: System Consider the case where there is dictionary data “/ about, reading: sei / yaku”.
[0053]
When the user selects a corresponding received mail (S201), an automatic word / character string extraction process is performed on the received mail (S202). At this time, in the sentence “Tomorrow will discuss the Doshu system” included in the received mail, it is a character string that follows the reading and precedes the word character string extraction keyword “About”, and the character type is “Doshu system” is extracted as a dictionary registration candidate as a character string composed of kanji. Here, by applying “about” with high accuracy as a word character string extraction keyword, the extraction accuracy is improved.
[0054]
Next, the original dictionary 21a is searched for the character unit reading of the extracted word character string. In this example, it is added as a condition that there are two or more candidates for the same reading as the word-by-character reading of the word character string. Then, “character: road, reading: how”, “character: state, reading: Shu”, “character: system, reading: se” are searched as “character string: road state, "Reading: Dosei" is created as a candidate for dictionary registration.
[0055]
According to the present embodiment configured as described above, in addition to the effects of the first embodiment, extraction of word character strings, generation of reading of word character strings, and reading of word character strings and readings by simply selecting document data The effect is that dictionary registration can be performed automatically.
In addition, the strictness of the word character string extraction conditions and the word character string reading and generation conditions can also prevent the accuracy of the registered dictionary data from being lowered.
[0056]
[Third embodiment]
Next, a third embodiment of the present invention will be described with reference to FIG.
FIG. 6 is a flowchart showing the operation of the word registration device according to the third embodiment of the present invention.
The third embodiment shown in this figure is different from the second embodiment in that the user sets selection conditions for document data without directly selecting the document data.
When the user inputs document data selection conditions (S301), word text string extraction processing (S302), reading generation processing (S303), and registration of word character strings and readings are performed on document data that meets the conditions. The processing (S304) is automatically performed, and then the reading can be converted into the word character string (S305).
[0057]
According to the third embodiment configured as described above, all word registration processing including document selection can be automated only by first setting the document selection condition. As a result, the user can perform Japanese conversion without being aware of unregistered words used in received mails, and full-scale context-dependent Japanese conversion processing is possible.
[0058]
【The invention's effect】
As described above, according to the present invention, a sentence that has already been converted into a character string is extracted, an unregistered word character string is extracted, its reading is generated, and the extracted word character string and the generated reading are read. By registering in the word dictionary, it is possible to reduce the word registration work by the user, and further simplify the word character string extraction processing and reading generation processing as much as possible, thereby limiting the hardware resources. It can also be implemented with a small communication terminal device such as the above.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a hardware configuration of a document input device (word registration device) according to a first embodiment of the present invention.
FIG. 2 is a block diagram showing a functional configuration of a document input device (word registration device) according to the first embodiment of the present invention.
FIG. 3 is a flowchart showing the operation of the word registration device according to the first embodiment of the present invention.
FIG. 4 is an explanatory diagram showing a word character string extraction keyword used by the word registration device according to the first embodiment of the present invention to extract a word character string.
FIG. 5 is a flowchart showing the operation of the word registration device according to the second embodiment of the present invention.
FIG. 6 is a flowchart showing the operation of the word registration device according to the third embodiment of the present invention.
[Explanation of symbols]
1 Document input device
2 Input section
3 Display section
4 Communication Department
5 storage unit
6 Control unit
10 Document database
20 Japanese translation device
21 word dictionary
21a Original dictionary
21b User dictionary
22 Conversion engine
30 word registration device
31 Word string extraction means
32 Reading generation means
33 word registration means

Claims (11)

A word registration device for storing a word character string and its reading in association with each other and registering an unregistered word character string and its reading for a word dictionary used for conversion from reading to word character string,
A word character string extraction means for extracting a word character string to be registered in the word dictionary using a word character string extraction keyword for a sentence that has already been converted into a character string;
The extracted word character string is decomposed into character units, readings of each decomposed character are searched from the word dictionary, and the plurality of searched readings are combined to generate the extracted readings of the word character string Reading generation means;
A word registration means for registering the extracted word character string and the generated reading in the word dictionary in association with each other;
With
The word character string extraction keyword consists of predetermined keywords including case particles, case particle equivalents, ramification particles, collection particles, case particles + collection particles, collection particles + case particles, connection particles, and judgment particles,
The word registration device, wherein the word character string extraction unit extracts a word character string to be registered in the word dictionary by extracting a word character string sandwiched between the word extraction keywords. The word character string extracting means extracts a word character string following a descriptive symbol and before the word extracting keyword, and among the extracted word character strings, those not registered in the word dictionary are registered. The word registration device according to claim 1, wherein: The word character string extracting means extracts a word character string composed only of kanji and / or a word character string composed only of katakana from a word character string extracted using the keyword for word extraction. The word registration device according to claim 1 or 2, characterized in that The word dictionary stores word character string readings separated in character character string units so that the word character string extracted by the word character string extracting unit is retrieved in character units. The word registration device according to claim 1.   The word dictionary includes an original dictionary and a user dictionary used for conversion from reading to word character string, and the reading generation means performs a reading search of the extracted word character string using the original dictionary, The word registration device according to claim 1, wherein the word registration unit performs word registration on the user dictionary.   The word registration device according to claim 1, wherein the word registration unit displays the extracted word character string and the generated reading, and requests correction and / or registration confirmation thereof. .   The word registration device according to claim 1, which is built in a small communication terminal device. The computer stores the word character string and its reading stored in the storage unit in association with each other, and stores the unregistered word character string and its reading with respect to the word dictionary used for conversion from reading to word character string. A word registration method to register,
Computer
For sentences already converted to strings, case particles, case particle equivalents, lemma particles, collection particles, case particles + collection particles, collection particles + case particles, connection particles, and judgment words stored in the memory A word character string to be registered in the word dictionary is extracted by extracting a word character string sandwiched between the word extraction keywords using a word character string extraction keyword including a predetermined keyword including:
The extracted word character string is decomposed into character units, the reading of each decomposed character is searched from the word dictionary, and a plurality of the searched readings are combined to generate an extracted reading of the word character string. ,
A word registration method, wherein the extracted word character string and the generated reading are associated and registered in the word dictionary.   The word registration method according to claim 8, which is performed in word registration in a small communication terminal device. Word for word registration device, is stored in association with the word string and its reading, for the word dictionary used in the conversion of readings Word string, to register the unregistered word string and read A registration program,
In the word registration device,
For sentences already converted to character strings, it consists of predetermined keywords including case particles, case particle equivalents, ramification particles, collection particles, case particles + collection particles, collection particles + case particles, connection particles, and judgment particles. Extracting a word character string to be registered in the word dictionary by extracting a word character string sandwiched between the word extraction keywords using a word character string extraction keyword,
The extracted word character string is decomposed into character units, the reading of each decomposed character is searched from the word dictionary, and the plurality of searched readings are combined to generate the extracted reading of the word character string. ,
A word registration program for associating the extracted word character string with the generated reading and registering it in the word dictionary. According to claim 10 word registration program for executing the word registration in small communication terminal device.

Images