JP2634596B2 - Kana-Kanji conversion device - Google Patents

Description

Description: BACKGROUND OF THE INVENTION The present invention relates to a kana-kanji conversion device for converting a kana input into a kana-kana-kana sentence, and more particularly, to a kana-kanji input which is not separated by a phrase unit. The present invention relates to an improvement of a kana-kanji conversion device that converts a sentence following an input without a special conversion start instruction means such as a conversion key.

[Conventional technology]

Conventionally, there are several known methods for converting a sentence of kana input (so-called solid input) which is not separated and written in units of bunsetsu into a sentence mixed with kanji kana.
There is a method described in Japanese Patent Application Laid-Open No. 60-189565, "Kana-Kanji conversion device".

According to the kana-kanji conversion device described in Japanese Patent Application Laid-Open No. 60-189565, kana input is accurately converted to a sentence mixed with kana-kana, and if the conversion result has ambiguity, a plurality of candidates are Can be extracted and held, and a correct conversion result can be selected from the extracted results.

Also, the method described in Japanese Patent Application Laid-Open No. 56-4832, "Input Method of Word Processor", automatically identifies the breakpoints of clauses and phrases based on the time interval of key input and converts them. This is a method in which conversion is automatically performed as timing, and there is no need for the user to press the conversion key and instruct conversion start.

Thus, of the conventional techniques described above, the former method is:
It is necessary to input a certain Kana sentence first, and then press the conversion key to instruct the start of Kana-Kanji conversion, and it takes time from pressing the conversion key to obtaining the conversion result,
In addition, even if the conversion method is not phrase-based, the operator must press the conversion key while being aware of the delimitation of the input sentence, which makes the operation cumbersome and changes the kana-kanji conversion candidate output on the display unit. Regarding the operation, it is necessary to always perform the partial change operation sequentially from the beginning of the character string of the kana-kanji conversion candidate.

In the latter example, the user does not need to recognize the conversion key, but since the system internally arbitrarily divides and converts the input reading sequence, it is not always possible to obtain the kana-kanji conversion result intended by the user. I can not say.

Japanese Patent Application Laid-Open No. Sho 62-97055 discloses that, when a text string intended by a user is not output, a partial change operation of the text string output on the display unit is always performed. There is disclosed a technique that does not need to be performed sequentially from the beginning of a text character string, but only needs to indicate a clause to be corrected.

The technique described in the above-mentioned Japanese Patent Application Laid-Open No. 62-97055 discloses a first storage means for storing a kana character string input from an input means, and a kanji / kana mixed character obtained by converting the kana character string to a kana / kanji character. A kana-kanji conversion device having a second storage means for storing a column, a character position in the first storage means is searched from an arbitrary character position in the second storage means, The sequence is re-converted into a Kanji-Kana-mixed character string at a delimiter position different from the phrase delimiter position of the Kanji-Kana-mixed character string in the second storage means.

As a specific example, for example, after converting a kana character string "sai-kankansori" to obtain a kanji-kana mixed sentence "reorganization / kansho / sato", re-conversion results in "re / conversion / processing". The case of correction will be described.

Note that the hatched “/” in the above and the following shows a segment break position.

First, it is assumed that the operator inputs a kana character string “saikankansori” from the keyboard and then instructs conversion, which is converted to “rearrangement / kansho / ri”. In response, the operator uses the cursor control keys to
As shown in " Edit / Kansho / Sato", the cursor is designated under "Edit" of the phrase "Reorganize" in which the segment break position is incorrect, and an instruction for re-conversion is issued. Accordingly, the control device activates the kana-kanji conversion unit using the reading “sai” in the heading character string of the kana-kanji dictionary as a key with a reading shorter than the reading of the phrase “sai-hen”, and converts the corresponding character “re”. After that, the remaining character string "refresh" is reconverted, thereby obtaining the character string "re / conversion / processing".

However, according to the technique described in Japanese Patent Application Laid-Open No. Sho 62-97055, upon re-conversion of a character string mixed with kanji and kana, an arbitrary character position in the second storage means is changed to a corresponding character position in the first storage means. The character position is searched, and the character string after that character is re-converted into a Kanji-Kana-mixed character string at a delimiter position different from the Kanji-Kana-mixed character string segmentation position in the second storage means. In addition to outputting a character string that changes the delimiter of a phrase, for a homonym, for example, the kana character string "saihenkanshori" in the above example, a process of storing fragments in a can is called "fragment / can". If you want to get the word "/ process", and if you first get "fragment / kansho / sato", you cannot get "fragment" which is a homonym of "reorganization". , The operator operates the converted kanji kana mixed character string There is such an inconvenience that it is necessary to temporarily convert the whole of "Kansho / Sato" back to the kana character string "Saikankansho" and perform Kana-Kanji conversion again to obtain the intended "Fragment / Can / Processing".

Japanese Patent Application Laid-Open No. Sho 62-11964 discloses a technique in which homonymous word correction and phrase segment change operation for a kana-kanji conversion result are performed by the same operation.

Also, Japanese Patent Application Laid-Open No. 61-190657 discloses a technique for automatically performing kana-kanji conversion from the leading end of the input reading in accordance with the input of the reading, based on the phrase separation information.

[Problems to be solved by the invention]

However, conventionally, the correction of the kana-kanji conversion result has been disclosed in Japanese Patent Application Laid-Open Nos. Sho 62-97055, 62-11964 and 61-61964.
Including Kana 190657 etc., always perform the correction operation on the Kana-Kanji conversion result immediately after inputting the reading and performing the first conversion operation. When the result requires further correction, a new candidate correction operation is required each time.

A first object of the present invention is to provide a kana-kanji automatic conversion device for automatically inputting a kana-kanji character string without pressing a special conversion key by inputting a kana character string which is not separated and written in units of phrases. After that, all kana-kanji conversion candidates were established.In the state before all the kana-kanji conversion operations were performed, after performing the kana-kanji conversion operation multiple times, that is, after performing all the kana-kanji conversion operations, the conversion results required further correction. In such a case, the conversion result can be corrected (hereinafter, referred to as “re-conversion” in the present invention) as easily as possible, and the operability at the time of inputting a document by Kana-Kanji conversion can be improved as compared with the conventional method. It is an object of the present invention to provide an improved kana-kanji conversion device capable of simultaneously changing a homonymous word when re-converting a mixed character string.

A second object of the present invention is to provide a device for inputting a kana character string which is not separated and written in units of phrases and automatically converting the kana to kanji characters without pressing a special conversion key. Establish all the Kana-Kanji conversion candidates after automatic conversion.In the state before all the Kana-Kanji conversion operations have been performed, after performing the Kana-Kanji conversion operation multiple times, that is, after performing the Kana-Kanji conversion operation once, further correction is required for the conversion result. In such a case, the conversion result can be corrected (re-converted) as easily as possible, and the operability during document input by Kana-Kanji conversion can be improved as compared to the past. It is an object of the present invention to provide a kana-kanji conversion device capable of changing both the case of changing a phrase break and the case of changing a homonym by the same operation.

[Means for solving the problem]

The first object is to provide an input unit for inputting a reading to be subjected to kana-kanji conversion, a unit for performing kana-kanji conversion without an instruction for kana-kanji conversion, a unit for dividing the kana-kanji conversion result into phrases, and an input character string. A storage unit for storing kana-kanji conversion candidates generated by the kana-kanji conversion process, a dictionary for storing information such as word notation and part of speech using word reading as index means, and extracting a partial character string from the input character string A kana-kanji conversion processing unit including a dictionary search unit for searching for a word having a corresponding reading in the dictionary, a kana-kanji conversion unit including a connection verification unit for preceding and following words based on the word searched by the dictionary search unit, A kana-kanji conversion device comprising: a display unit for displaying a more input character or a kana-kanji conversion result; and a means for establishing a kana-kanji conversion target character string while selecting and displaying it. A memory for storing an input character string and a character string of a Kana-Kanji conversion result in association with a phrase unit, means for correcting and instructing the Kana-Kanji conversion result, and establishing all Kana-Kanji conversion candidates after the automatic conversion of Kana-Kanji characters. In the state before all are established, move the cursor to the place you want to correct and re-convert any phrase,
Achieved by adding means for outputting homonyms and correcting kana-kanji conversion results upon re-conversion, and means for rewriting the memory so as to correspond to the phrase conversion results corrected by the correction operation. Is done.

Further, the second object is to provide an input section for inputting a reading to be subjected to kana-kanji conversion, means for performing kana-kanji conversion without an instruction for kana-kanji conversion, means for dividing the kana-kanji conversion result into phrases, and A storage unit for storing kana-kanji conversion candidates generated by the character string and kana-kanji conversion processing, and word notation,
A dictionary that stores information such as part of speech, a dictionary search unit that extracts a partial character string from the input character string and searches for a word having a corresponding reading in the dictionary, and a word that is searched by the dictionary search unit. A kana-kanji conversion processing unit including a connection verification unit for preceding and succeeding words, a display unit for displaying a character or a kana-kanji conversion result input from the input unit, and a unit for establishing a kana-kanji conversion target character string while displaying and selecting the character string. A kana-kanji conversion device comprising: a memory for storing an input character string and a character string of a kana-kanji conversion result in association with a phrase unit; a means for instructing to correct the kana-kanji conversion result; and The phrase information before the phrase stored in the memory is added, and after the kana-kanji automatic conversion, all kana-kanji conversion candidates are established. Move the cursor to a new position and re-convert any phrase.In the re-transformation, change both the case of changing the segment break and the case of changing the homonym by the same operation. This is achieved by adding means for outputting a corrected character string converted into Kana-Kanji characters, and means for rewriting the memory so as to correspond to a new phrase conversion result corrected by the correction operation.

[Action]

According to the first invention, a memory for storing an input character string and a character string of a kana-kanji conversion result in association with a phrase unit, a unit for instructing a correction of the kana-kanji conversion result, and after kana-kanji automatic conversion, Establish all of the Kana-Kanji conversion candidates Move the cursor to the point you want to correct before re-establishing all the words and re-convert any phrase, and output the same-synonym to correct the Kana-Kanji conversion result Means, and a means for rewriting the memory so as to correspond to the result of the conversion of the phrase corrected by the correction operation. Establishes all kana-kanji conversion candidates after automatic kana-kanji conversion in a device that automatically converts kana-kanji characters without pressing a special conversion key In the previous state, after performing the Kana-Kanji conversion operation multiple times, that is, after performing the Kana-Kanji conversion operation once, if the conversion result requires further correction, simply specify the phrase to be corrected. ,
Once again, a matrix display of conversion candidates appears, and you can select the target candidate from the selection matrix. Correction (re-conversion) of the conversion result is made as simple as possible. You can also change the synonym.

Further, the second invention is a memory for storing the input character string and the character string of the kana-kanji conversion result in association with each phrase unit, and means for instructing the kana-kanji conversion result to be corrected,
In the state before adding all the kana-kanji conversion candidates, the kana-kanji automatic conversion is added, and the kana-kanji conversion candidate is added after adding the phrase information before the phrase stored in the memory by the correction instruction means.
Move the cursor to the location you want to modify and re-convert any phrase.At the time of re-translation, change both the phrase break change and the homonym change with the same operation, Means for outputting a corrected character string converted to kana-kanji, and means for rewriting the memory so as to correspond to the result of conversion of a new phrase corrected by the correction operation, so that writing is performed in units of phrases. In a device that automatically converts Kana-Kanji characters without inputting a special Kana-Kanji character string without pressing a special conversion key, establishes all Kana-Kanji conversion candidates after Kana-Kanji automatic conversion. Then, after performing the Kana-Kanji conversion operation multiple times, that is, after performing the Kana-Kanji conversion operation once, if the conversion result requires further correction , Simply instructs the clause to be a correction, it appeared again, matrix display comprising a conversion candidates,
You only have to select the target candidate from the selection matrix, and the conversion result can be corrected (re-converted) as easily as possible. In addition, when re-converting a character string mixed with kanji or kana, multiple selections output in a matrix format on the display unit A possible text character string is a character string output by performing Kana-Kanji conversion based on the two phrase information of the phrase containing the character string instructed to be corrected and the phrase before it. It is a text string in which a text string to be changed and a text string to be a homonym are mixed. Therefore, the user can change both the case of changing the segment break and the case of changing the homonym by the same operation when re-converting the character string mixed with kanji and kana.

〔Example〕

Hereinafter, the present invention will be described with reference to one embodiment of the drawings. FIG. 2 is an overall hardware configuration diagram of a kana-kanji conversion device. In FIG.
Denotes a display unit, 2 denotes a control unit, 5 denotes a storage device, and 3 denotes an external storage device.

In the above configuration, the input unit 1 is an input device for inputting an input reading sequence, kana-kanji conversion, a selection instruction, and the like, and the display unit 4 displays input readings, and further displays kana-kanji conversion results and the like. For example, it is constituted by a CRT display. The control unit 2 controls execution of kana-kanji conversion, and is constituted by a CPU. The storage device 5 is a place where programs and data are stored, and the external storage device 3 is a place where a dictionary used for kana-kanji conversion is stored.

FIG. 1 is a flowchart showing the entire operation system of the kana-kanji conversion process by the apparatus of the present invention. Hereinafter, the contents of the kana-kanji conversion process will be described with reference to FIG.

In the key input analysis process (step 10 in FIG. 1), the input information from the input unit 1 is analyzed to handle the input information, such as performing a kana-kanji conversion process and creating a character string for candidate correction. Perform the processing to be performed.

If the character string of the Kana-Kanji conversion result is output as a result of the key input analysis processing (10), the conversion result character string processing (20)
And displays a character string of the conversion result (hereinafter, referred to as a text character string) on the display unit 4 and an input reading management buffer of the storage device 5 so that the input reading and the text character string correspond to each phrase unit. And stored in the text string management buffer.

If the result of the key input analysis process (10) is a conversion key,
The input reading and text string of the phrase including the text character string indicated by the cursor and the input reading and text string of the phrase before the phrase are extracted from the input reading management buffer and the text character string management buffer of the storage device 5. Is passed to the selection candidate creation processing (40).

In the selection candidate creation process (40), the kana-kanji conversion process is performed by adding the input reading of the phrase pointed to by the cursor to the input reading of the previous phrase, and the information including the information of the previous phrase from the beginning of the phrase pointed by the cursor is corrected. Outputs multiple possible text strings.

If the result of the key input analysis processing (10) is a correction of a conversion candidate, the candidate correction processing (50) is started, and the text character string already displayed on the display unit 4 is replaced with the text character string selected by the user. Change to

If the result of the key input analysis processing (10) establishes a part of the phrase that is being corrected, the partial establishment processing (60)
Is started, and a part of the text string being corrected displayed on the display unit 4 is established.

If the result of the key input analysis process (10) establishes all of the clauses that are being corrected, the all-establishment process (70) is started and the text character string displayed on the display unit 4 is activated. Establish all of

Finally, if the result of the key input analysis processing (10) is the end key, the document editing is terminated, otherwise, the key input analysis processing (10) is started again.

The key input analysis processing (10), the conversion result character string processing (20), the input reading character string extraction processing (30),
Each operation of the selection candidate creation process (40), candidate correction process (50), partial establishment process (60), and all establishment process (70) will be described in detail.

FIG. 4 is a flowchart showing an operation system of the key input analysis processing (10).

In FIG. 4, first, at step 105, it is determined whether the input key is retracted or not. The input key is saved at a timing other than the conversion key (902), the change key (904), and the partial establishment key (906) shown in FIG. 3 while the text character string for correction is being output. Is entered. The reason is that it is necessary to output information for establishing a text string and to establish all text strings to be corrected. And whether the input key is retracted,
The control of whether or not there is is performed by FLG2. The input key is FLG2
Is not saved when is "0", is saved when FLG2 is "1", and when FLG2 is "0", in step 110, information input by the user is extracted from the input unit 1 in step 110, and When "1" is set, the input information is extracted from the save area and stored in the memory of the storage device 5 (hereinafter referred to as INPUT @ KEY).

In step 115, FLG1 is used to determine whether a certain clause in the text string is in a corrected state by outputting a text string for correction or in a state where a normal document is being created. Use to judge. When FLG1 is "1", it is in the correction state, when FLG1 is "0", it is in the document creation state, and FLG1 and FLG2 are cleared to "0" as initial settings when the system is started.

If FLG1 is “0” in step 115, that is, if a document is being created,
It is determined whether NPUT @ KEY is a character key (901) in FIG. In the case of the character key (901), in step 130, a character input process is performed.

FIG. 5 is a flowchart showing an operation flow system of the character input processing (130).

In FIG. 5, first, at step 1310, IN
Input reading storage area for PUT ＄ KEY information (hereinafter CHR ＄ B
UFF). At this time, CHR ＄ POINTER is used as a pointer for managing CHR ＄ BUFF, and CHR ＄ P
OINTER sets “1” when the system is started up. In step 1320, CHR１POINTER is incremented by one. In step 1330, the input reading character string stored in CHR ＄ BUFF is displayed on the display unit 4. indicate. For example, the sixth
As shown in the figure, when the user inputs the character key “?” From the input unit 1 in a state where “Singaku” is stored in the CHR ＄ BUFF, in step 1310, the “?”
It is stored as shown in the figure, and in step 1330, it is displayed on the display unit 4 as shown in FIG. In FIG. 8, the lowermost part of the screen shows an input line, and the upper part shows a text character string output line.

In FIG. 4, after completing the character input processing (130),
In step 140, it is checked whether the Kana-Kanji conversion process can be started with the current input character string.

FIG. 9 is a flowchart showing the operation system of the kana-kanji conversion activation check.

In FIG. 9, first, at step 1410, IN
PUT ＄ KEY determines whether it is a punctuation mark. In the case of a punctuation mark, "1" is set to a conversion flag instructing the start of conversion. If it is not a punctuation mark, it is determined whether or not it is a break between phrases (1420). If it is determined that it is a break between phrases, the conversion flag is set to "1" (1430). “0” is set (1440).

In FIG. 4, in step 142, it is determined whether or not the conversion flag is "1". If the conversion flag is "1", the conversion process (80) is started to perform kana-kanji conversion. If it is "0", the next input key is taken out (110).

In the conversion process (80) of FIG. 4, based on the input reading stored in CHR ＄ BUFF, a dictionary having word reading, notation, and part of speech is searched in order from the beginning, and the searched word has Using the part of speech, the operation of performing a connection test with the preceding and following words is repeated to create kana-kanji conversion candidates. For example,
FIG. 11 shows a specific example of data formed as kana-kanji conversion candidates in CHR ＄ BUFF shown in FIG. 10, and the conversion candidates adopt a network-like data structure in order to prevent duplication of data. Inside, the notation, the corresponding reading and the part of speech are stored, thereby reducing the storage capacity. As a method of selecting an optimal candidate from the created network, for example, as shown in FIG. 5 of JP-A-60-189565, a candidate string having the highest likelihood is selected from the candidate strings. It is extracted and stored as a conversion result in a conversion candidate display buffer (hereinafter referred to as HENKAN @ BUFF). Here, the likelihood means that the number of phrases is obtained by multiplying the weight in consideration of the part of speech and the appearance frequency, and the smaller the number is, the higher the likelihood is. Specifically, the likelihood is noun, verb, Adjectives and adjective verbs are given a weight of 1, formal nouns, auxiliary verbs, adnominals, etc. are given a weight of 0.1, and prefixes and suffixes are given a quasi-independent word, giving a weight of 0.5. As shown in Fig. 12, HENKAN
格納 Store in BUFF.

In FIG. 4, in step 150, the text string stored in HENKAN @ BUFF is divided into segments based on the parts of speech stored in the data structure on the network, and as shown in FIG. The number is stored in the number of valid phrases, the number of input readings of each phrase is stored in the input reading number management area of each phrase, and the number of text character strings is stored in the text character string number management area of each phrase. In the case of the example shown in FIG. 12, since there are two clauses of “mathematics” and “analytical”, “2” is set as the number of effective clauses,
In the input reading count management area (1) for each clause,
In the input reading count management area (2), the number of characters of the input reading “sugaraku” corresponding to
The number of characters “8” of “Kyosei Shikino” is stored. The number of characters “2” of “mathematics” is stored in the text character string number management area (1) per phrase, and the number of characters “analytical” is stored in the text character string number management area (2) per phrase. 4 "is stored.

In FIG. 4, the conversion process (80) is started by the
This is also performed when the conversion key (902) shown in FIG. 3 is input while the input reading is stored in R ＄ BUFF. In addition, when the input of the conversion key (902) is not stored in the CHR @ BUFF, the information of the conversion key is output.

Next, the conversion result character string processing (20) started when the output of the key input analysis processing (10) in FIG. 1 is a character string of the conversion result will be described with reference to FIG.

In FIG. 1, in the conversion result character string processing (20), the number of valid phrases set in the output character string creation (step 150 in FIG. 4), the input reading count management area for each phrase, the text character string for each phrase Based on the number management area, as shown in Fig. 14, the input reading from CHR ＄ BUFF is read from HENKAN ＄ BUFF.
, Take the text strings from
The text character strings are stored in the input reading management buffer in the storage device 5 in correspondence with the text character string management buffers in the storage device 5, respectively. Thus, specific examples of the input reading management buffer and the text character string management buffer are shown in FIG. 15 and FIG. 16, and in FIG. 15 and FIG.
"/" Was written. FIG. 17 shows an example of the display state of the display unit 4 in the output of a text string, and converts the input reading into a text string according to the flow shown in FIG. It is displayed on the display unit 4.

Here, a flow of correction when the text character string displayed on the display unit 4 is different from the text character string intended by the user will be described.

After moving the cursor to the text character string to be corrected, the user inputs a conversion key (902) shown in FIG. In the key input analysis processing (10) in FIG. 1, it is checked whether or not an input reading exists in CHR @ BUFF based on step 145 in FIG. If there is no input reading in CHR ＄ BUFF, a conversion key 902 is output. After the determination in step 25 in FIG. 1, the input reading character string extracting process (30) is activated to extract the input reading corresponding to the phrase including the text character string to be corrected.

FIG. 18 is a flowchart showing an operation system of the input reading process (30).

In FIG. 18, first, in step 310, it is checked how many words the cursor position is from the head of the text character string management buffer, and in step 315, the corresponding input reading is extracted from the input reading management buffer. Then, the input reading extracted from the input reading management buffer is
Set to SUB1 \ CHR \ BUFF (320). The number of characters set is set in SUB1 \ CHR \ CNT (325). Similarly, in step 330, the text string is extracted from the text string management buffer, and in step 335, SUB1 ＄ TEXT ＄ BUFF is set, and in step 340, the number set in SUB1 ＄ TEXTSUBBUFF is set to SUB1 ＄ TEXT. ＄ C
Set to NT.

Next, in step 345, it is checked whether the phrase before the phrase to be corrected is a phrase holding the input reading, and if the input reading exists, the reading is taken out from the input reading management buffer. (350), SUB2 ＄ CHR ＄ BU
Set to FF (355). Then, the number of characters set in SUB2 ＄ CHR ＄ BUFF is set in SUB2 ＄ CHR ＄ CNT (36
0). Similarly, extract the text string and use SUB2 ＄ TEXT
Set in ＄ BUFF and SUB2 ＄ TEXT ＄ CNT (365,370,3
75). If the previous clause does not exist and does not hold the input reading, SUB2 ＄ CHR ＄ CNT and SUB2 ＄ TEXT ＄
"0" is set to CNT (380). In steps 385 and 390, SUB1 ＄ is entered in the input reading length area and the text character string length area used in the process of establishing a corrected phrase.
Set the value of CHR ＄ CNT and the value of SUB1 ＄ TEXT ＄ CNT as initial values. Also, in step 395, a SET for managing the input reading length area and the text character string length area
「Set“ 1 ”to CNT. For example, in the state where the text string as shown in FIG. 19 is output on the display unit 4 in FIG. 2, after moving the cursor to the position of “kai”, the user inputs the conversion key (902) in FIG. In the case, from the input reading management buffer shown in FIG. 20 and the text character string management buffer shown in FIG. 21, the input reading and text characters corresponding to the phrase "on the table" at the cursor position and the previous phrase "mathematics" The column is taken out and information is stored in each memory of the storage device 5 as shown in FIG.

In the input reading character string extracting process (30) of FIG. 1, after the input character string is extracted, the selection candidate creating process (40)
Invokes and outputs a modifiable text string.

FIG. 23 is a flowchart showing an operation system of the selection candidate creation processing (40).

In FIG. 23, in steps 410 and 420,
SUB2 ＄ CHR ＄ BUFF and SUB1 ＄ CHR ＄ BUFF are set in CHR ＄ BUFF, and input reading for starting the conversion process (80) is created. In the process (80), a conversion process is performed based on CHR ＄ BUFF to create a network-like data structure as shown in FIG. In step 430, it is determined whether or not a character string identical to the text character string stored in SUB2 \ TEXT \ BUFF exists at a corresponding position in the created network-like data structure. Step 43
At 0, if there is a character string identical to the text string stored in SUB2 ＄ TEXT ＄ BUFF, it is followed by the same text string stored in SUB1 ＄ TEXT ＄ BUFF Check if the character string exists (45
0). In step 460, if the same character string as the text string stored in SUB1 ＄ TEXT ＄ BUFF does not exist, in step 470, the contents of SUB1 ＄ CHR ＄ BUFF are reset to CHR ＄ BUFF, Restart the conversion process (80). Then, in step 460, SUB1 ＄ TEXT ＄ B
If the same character string as the text string stored in UFF follows, a plurality of text strings to be connected after the previous clause are added, and in step 460, SUB1 ＄ TEX
If the same character string as the text string stored in T ＄ BUFF does not follow, multiple text strings starting from the beginning are output in a matrix format, for example, as shown in the lower right of FIG. 24. (Step 480). And the second
As shown in Fig. 4, the output multiple selectable text strings perform kana-kanji conversion based on the two pieces of phrase information, the phrase containing the character string instructed to be modified and the phrase before it. Since this is a character string output as a text string, the text string is a text string in which a text string for changing a phrase break and a text string for changing a homonym are mixed. Therefore, the user can change both the case of changing the segment break and the case of changing the homonym by the same operation.

Next, a plurality of candidate text strings that can be corrected are
Described below is a case where a change is made using the change key (904) shown in the figure.

In FIG. 1, after activating the selection candidate creation process (40) and outputting a plurality of correctable text character strings, it is recognized in step 45 that "1 And set the key input analysis process (10)
Start The user operates the change key (904) shown in FIG.
Accordingly, when a character string intended by the user is selected from a plurality of correctable text character strings, a changed character string creation processing system (160) shown in FIG. 4 is started.

FIG. 25 is a flowchart showing an operation system of the modified character string creation processing (160).

In FIG. 25, first, in step 1605, the optimum character string following the selected text character string is extracted from the network-like data structure, and is set to HENKAN ＄ BUFF, including the selected character string. (1610). next,
The text string set in HENKAN @ BUFF is divided into clauses, and the length of the input reading and the length of the text string corresponding to each clause are stored in the input reading length area and the text string length area. In addition, set the number of segments
To be stored. For example, when “family” is selected in the state of FIG. 24, as shown in FIG. 26, HENKAN ＄ BUFF
"Department transfer" is retrieved from the network-like data structure and stored.
The information is stored in the input reading length area, the text character string length area, and SET @ CNT. And
In the candidate correction process (50) shown in FIG.
The text string in the BUFF is displayed on the display unit 4 as shown in FIG. In the state of FIG. 27, when the user inputs the partial establishment key (906) of FIG. 3, the user starts the partial establishment character string creation (170) of FIG. , “Family” is established and displayed on the display unit 4. That is, comparing FIG. 27 with FIG. 34, it can be seen that “family” is outside the rectangular frame in FIG.

FIG. 28 is a flowchart showing an operation system for a character string creation process (170) in which a text character string is partially established.

In FIG. 28, in step 1705, it is checked whether or not the text string to be established is a part of the preceding phrase. If the text string to be established is a prefix or an independent word without a prefix, “0” is set in the phrase information area as an independent phrase, and the text string to be established is an independent word with a prefix, In the case of a word or an adjunct, "1" is set in the phrase information area as a part of the preceding phrase. An independent word with a prefix here refers to a case where the text string to be established is an independent word and the text string immediately before it is a prefix. Next, the input reading length area and the text character string length area are corrected (1765, 1770). The correction is to subtract the number of input readings and the number of text strings corresponding to the character string to be established from (1) of each area (1745), and if (1) of each area becomes "0", , Of each area (2)
Is moved to (1) and (3) to (2). Also, one is subtracted from SET @ CNT (1760). In the example of FIG. 26, since the input reading length of “Family” is 1 and the text string length is 1, when the “Family” is established, the input reading length area (1) and the text character string length area (1) becomes “0”. Therefore, (2) of each area is moved to (1), and finally, “7” is entered in the input reading length area (1), “4” is entered in the text character string length area (1), and SET ＄ "1" for CNT
Is stored.

FIG. 29 is a flowchart showing the operation system of the text character string part establishing process indicated by reference numeral 60 in FIG.

In FIG. 29, in steps 605 and 610, the input reading management buffer and the text character string management buffer are segmented. For example, when the "family" is established in the state of FIG. 27, as shown in FIG. 30 and FIG. 31, the "class" is divided into "ka" and "class". Next, in step 615, if the phrase information area is "1", the process is connected to the previous phrase. For example, in the case of "family", since it is a suffix of "mathematics" of the previous phrase, the phrase information area is "1", and as shown in FIG. 32 and FIG. "", And create a new phrase "Sugarakuka".

Finally, the case where the end key (908) in FIG. 3 is input to establish all the text strings will be described. That is,
FIG. 35 is a flow chart showing the operation system of the process for establishing the entire text character string indicated by reference numeral 70 in FIG.

In FIG. 35, in step 705, SET ＄ CNT
The input reading management buffer is updated based on the input reading length area. In step 710, the text character string management buffer is updated by the number of SET / CNT based on the text character string length area. Next, in the process 715, the part of speech of the leading established character string is examined, and in the case of an independent word with a prefix, a suffix, or an adjunct, in the processes 735 and 740, the input reading length area (1) and the text character The character string corresponding to the column length area (1) is connected to the previous phrase.

The kana-kanji conversion device according to the present invention has a configuration in which kana character strings that are sequentially input are sequentially converted to kana-kanji characters, and the kana-kanji character strings are sequentially input as in an interactive voice input device. It is also effective for various devices.

〔The invention's effect〕

The present invention is as described above, and the present invention (first invention)
Is a memory for storing an input character string and a character string of a kana-kanji conversion result in association with a phrase unit, means for instructing correction of the kana-kanji conversion result, and establishing all kana-kanji conversion candidates after automatic kana-kanji conversion. In a state before all of them are established, the cursor is moved to a position to be corrected to re-convert any phrase, and at the time of the re-conversion, a homonym is output to correct the kana-kanji conversion result, and the correction operation is performed. Means for rewriting the memory so as to correspond to the conversion result of the phrase corrected by the above, inputting a character string which is not divided and written in units of a phrase, and pressing a special conversion key Without automatic conversion of Kana-Kanji characters, Kana-Kanji conversion is automatically performed after all Kana-Kanji conversion candidates are established. After performing several times the conversion operation,
In other words, after performing the kana-kanji conversion operation once, if the conversion result requires further correction, simply indicating the phrase to be corrected, the matrix display as a conversion candidate appears again, and from the selection matrix It is only necessary to select the target candidate, the conversion result can be corrected (re-converted) as easily as possible, the operability during document input by Kana-Kanji conversion can be improved compared to the past, and the character string mixed with Kana-Kana At the time of the re-conversion, as shown in the uppermost row of the lower right matrix candidate in FIGS. 24 and 27, for example, the homonym change selection can be simultaneously performed.

Also, the present invention (second invention) provides a memory for storing an input character string and a character string of a Kana-Kanji conversion result in association with each phrase, a means for instructing the Kana-Kanji conversion result, Means for adding the phrase information before the phrase stored in the memory, and after automatic kana-kanji conversion, establishing all kana-kanji conversion candidates. Means for re-converting and re-converting, in the same operation, both the case of changing a phrase break and the case of changing a homonymous word, and outputting a corrected character string converted into a new kana-kanji character. And means for rewriting the memory so as to correspond to the result of the conversion of the new phrase corrected by the correction operation described above. Input a character string and automatically convert it to a Kana-Kanji character without pressing a special conversion key. After automatic Kana-Kanji conversion, establish all Kana-Kanji conversion candidates. After performing the operation multiple times, that is, after performing the kana-kanji conversion operation once, if the conversion result requires further correction, simply specify the phrase to be corrected and become a conversion candidate again A matrix display appears, and you can select the target candidate from the selection matrix. Correction (re-conversion) of the conversion result can be performed as easily as possible, and operability at the time of document input by Kana-Kanji conversion can be improved compared to the past. In addition, when re-converting a character string mixed with kanji and kana, a plurality of selectable text strings output in a matrix format on the display unit are A character string output by performing Kana-Kanji conversion based on the two phrase information of the phrase including the character string designated as positive and the preceding phrase, , And a text character string in which a homonym is changed. Therefore, when re-converting a character string mixed with kanji or kana,
As shown in FIG. 27 and the lower right matrix candidate all column in FIG. 27, it is possible to change both the case of changing the segment break and the case of changing the homonym by the same operation.

[Brief description of the drawings]

1 shows an embodiment of a kana-kanji conversion device according to the present invention, FIG. 1 is a flowchart showing an overall operation system of a kana-kanji conversion process by the device of the present invention, FIG. 2 is an overall hardware configuration diagram of the device of the present invention, FIG. 3 is a key layout diagram, FIG. 4 is a flowchart showing an operation system of key input analysis processing, FIG. 5 is a flowchart showing an operation system of character input processing, and FIG.
FIGS. 7, 7, and 10 show specific examples of the input reading storage area, respectively, and FIGS. 8, 17, 19, 24, and 27.
FIG. 34 and FIG. 34 each show a specific example of screen output information.
9 is a flowchart showing an operation system of a kana-kanji conversion activation check, FIG. 11 is a diagram showing a specific example of data formed by a kana-kanji conversion unit, and FIG. 12 is a diagram showing a specific example of a kana-kanji conversion candidate display buffer. FIG. 13 is a diagram showing a specific example of a management area when the kana-kanji conversion candidate display buffer is divided into phrases, FIG. 14 is a flowchart showing an operation system of kana-kanji conversion result character string processing, and FIGS. Figure, Figure 30,
FIG. 32 is a diagram showing a specific example of an input reading management buffer, and FIG.
FIGS. 21, 21, 31 and 33 show specific examples of the text character string management buffer, FIG. 18 is a flowchart showing the operation system of the input read character string retrieval processing, and FIG. A diagram showing a specific example of a memory for creating a plurality of character strings, FIG.
FIG. 23 is a flowchart showing an operation system of the selection candidate creation process, FIG. 25 is a flowchart showing an operation system of the changed character string creation process, and FIG. 26 is a management area when the kana-kanji conversion candidate display buffer is divided into segments. FIG. 28 shows a specific example.
FIG. 29 is a flowchart showing an operation system for a character string creation process in which a text character string is partially established. FIG. 29 is a flowchart showing an operation system for a text character string portion establishment process.
The figure is a flowchart showing the operation system of the entire text character string establishment process. 1 input unit 2 control unit 3 external storage device 4
... Display unit, 5 ... Storage device.

 ──────────────────────────────────────────────────続 き Continuation of the front page (72) Inventor Takuya Watanabe 1-1-1, Higashitaga-cho, Hitachi City Inside the Taga Plant of Hitachi, Ltd. (56) References JP-A-56-38674 (JP, A) 60-189565 (JP, A) JP-A-62-97055 (JP, A)

Claims (2)

(57) [Claims] 1. An input unit for inputting a reading to be converted to kana-kanji, means for performing kana-kanji conversion without an instruction for kana-kanji conversion, means for dividing the kana-kanji conversion result into phrases, an input character string and kana-kanji A storage unit that stores kana-kanji conversion candidates generated by the conversion process, a dictionary that stores information such as word notation and part of speech using word reading as indexing means, and a partial character string cut out from the input character string, A dictionary search unit for searching for a word having a corresponding reading in the dictionary, a kana-kanji conversion processing unit including a connection verification unit for preceding and following words based on the word searched by the dictionary search unit, and an input from the input unit Kana-Kanji conversion device equipped with a display unit for displaying the converted characters and Kana-Kanji conversion results, and a means for establishing the Kana-Kanji conversion target character string while displaying and selecting it. A memory for storing the input character string and the character string of the kana-kanji conversion result in association with each phrase unit, means for instructing the kana-kanji conversion result to be corrected, and establishing all kana-kanji conversion candidates after automatic kana-kanji conversion. In a state before the whole is established, the cursor is moved to a place not to be corrected, and an arbitrary phrase is re-converted, and at the time of the re-conversion, a homonym is output to correct the kana-kanji conversion result, and the correction operation is performed. A rewriting means for rewriting the memory so as to correspond to the phrase conversion result corrected by the method. 2. An input unit for inputting a reading to be subjected to kana-kanji conversion, means for performing kana-kanji conversion without an instruction for kana-kanji conversion, means for dividing the kana-kanji conversion result into phrases, an input character string and kana-kanji characters A storage unit that stores kana-kanji conversion candidates generated by the conversion process, a dictionary that stores information such as word notation and part of speech using word reading as indexing means, and a partial character string cut out from the input character string, A dictionary search unit for searching for a word having a corresponding reading in the dictionary, a kana-kanji conversion processing unit including a connection verification unit for preceding and following words based on the word searched by the dictionary search unit, and an input from the input unit Kana-Kanji conversion device equipped with a display unit for displaying the converted characters and Kana-Kanji conversion results, and a means for establishing the Kana-Kanji conversion target character string while displaying and selecting it. A memory that stores the input character string and the character string of the kana-kanji conversion result in units of phrases, a unit that instructs the kana-kanji conversion result to be modified, and a unit that corrects the phrase stored in the memory by the modification instruction unit. After adding the previous phrase information and automatically converting Kana-Kanji characters, establish all of the Kana-Kanji conversion candidates.In the state before all the kana-kanji conversion is established, move the cursor to the position to be corrected, re-convert any phrase, and perform the re-conversion. Means for changing the phrase segmentation and for changing the homonymous words by the same operation, and outputting a new kana-kanji converted corrected character string; and A kana-kanji conversion device, further comprising means for rewriting the memory so as to correspond to a conversion result of a new phrase.