JPH02299060A - Kana-kanji converting device - Google Patents

Abstract

PURPOSE:To accurately segment paragraphs by reading a KANA(Japanese syllabary) string corresponding to a KANA-KANJI(Chinese character) mixed character string retrieved by a retrieval means, out of a determined paragraph storage means, and reconverting the KANA character string through a conversion dictionary storage means. CONSTITUTION:When a conversion key 2b is operated, a conversion part 10 refers to a learning memory 9 and performs a conversion analysis for the KANA-KANJI conversion of a KANA character string in an input buffer 7. Then the memory 9 judges whether or not the KANA-KANJI conversion is possible and when the conversion is possible, a conversion candidate which is obtained by the conversion is stored in a candidate buffer 11 and displayed on a display part 3. An operator judges its display contents and performs key operation for indicating various functions. When a determination key 2d is operated, the conversion candidate in the buffer 11 is stored in a text memory 4 and the conversion contents are determined. Then the conversion candidate which is converted and determined is made to correspond to its reading of KANA in paragraph units and stored on the memory 9, the contents of the buffers 7 and 11 are cleared, and the conversion is finished.

Description

[Detailed Description of the Invention] [Field of Industrial Application] The present invention relates to a kana-kanji conversion device that converts an input kana character string into document data of a mixed kana-kanji character string, and in particular, relates to a kana-kanji conversion device that converts an input kana character string into document data of a mixed kana-kanji character string, and particularly relates to a document data whose conversion has been confirmed. The present invention relates to a kana-kanji conversion device suitable for re-converting data.

[Prior art] Conventionally, an input kana character string that has been confirmed to be converted into document data consisting of a mixed kana/kanji character string is saved as is even after the conversion is confirmed, and this is used to convert it into another kana character string again. A kana-kanji conversion device that reconverts into a kanji-mixed character string has been realized.

[Problem to be solved by the invention] However, when specifying a character string to be reconverted for a kana-kanji mixed character string whose conversion has been confirmed, the operator had to accurately specify the segment break position. .

This is because in conventional reconversion, the kana character is extracted from the position of the stored kana character that corresponds to the position specified by the operator, and the phrase is extracted by grammatical analysis such as the longest match method. If you specify the wrong position when
This is because completely meaningless grammatical analysis is performed.

An object of this invention is to enable, even if an arbitrary position on fixed document data is specified, a clause containing a character at this position to be accurately extracted and reconversion to be performed accurately within the range of this clause. It is.

[Means to solve the problem] The means of this invention are as follows.

This kana-kanji conversion device converts input kana character strings into phrases based on a conversion dictionary storage means a (see the functional block diagram in FIG. 1, the same applies hereinafter) that stores kanji and their pronunciations in correspondence unit is converted into a kana-kanji mixed character string, and by determining the kana-kanji mixed character string of each clause, each kana-kanji mixed character string is written as document data into the document storage means, and the conversion dictionary storage means In addition to the document storage means a and the document storage means, it has a fixed phrase storage means C1, a retrieval means d, and a reconversion means e.

The determined phrase storage means C stores the kana-kanji mixed character string of each determined phrase and the corresponding kana character string in association with each other.

When an arbitrary character position on the document data in the document storage means is specified under the reconversion command, the search means d searches for a character string that matches the kana-kanji mixed character string before and after the specified character position. is searched by comparison with each kana-kanji mixed character string stored in the confirmed phrase storage means C.

The re-conversion means e reads out the kana character string corresponding to the kana-kanji mixed character string searched by the search means d from the confirmed phrase storage means C, and re-converts the kana character string based on the conversion dictionary storage means ac: .

[Effect] The operation of the means of this invention is as follows.

The kana character string "This year's thickness is" that has just been inputted is converted into the kana-kanji mixed character string "This year's thickness is" on a bunsetsu basis, and the confirmation instruction for "This year's thickness is" is executed. It shall be assumed that

Then, "This year's thickness is" is written into the document storage means C as document data, but at this time, the fixed phrase storage means C
stores the confirmed kana-kanji mixed character string and the corresponding kana character string in association with each other in clause units. In other words, "this year's" is remembered in correspondence with "this year's", and "
``Thickness wa'' is stored in correspondence with ``Atsusa wa''.

It is assumed that, under the reconversion command, for example, "sa" in the document data "thickness" in the document storage means is arbitrarily designated. Then, the search means d searches the determined phrase storage means C for "Tsukuwa" as a character string that matches the kana-kanji mixed character strings before and after the designated character position. In addition, since the fixed phrase storage means C stores kana-kanji mixed character strings (conversion contents) and the corresponding kana character strings for each phrase as described above, "thickness is". Similarly, if you specify "thickness" or "wa",
"Thickness is" is searched.

Then, the re-conversion means e reads out the kana character string "Atsusaha" corresponding to the kana-kanji mixed character string "Atsushi wa" searched by the search means d from the determined phrase storage means C, and `` is reconverted into ``heat is j, etc.'' based on the conversion dictionary storage means a.

Therefore, even if an arbitrary position on the determined document data is specified, a clause containing a character at this position can be accurately extracted, and accurate reconversion can be performed within the range of this clause.

[Example] An example will be described below with reference to FIGS. 2 to 6.

FIG. 2 is a block diagram of the kana-kanji conversion device.
The CPU converts the kana character string inputted from the key human power unit 2 into a kana-kanji mixed character string in units of clauses, displays the conversion results as conversion candidates on the display unit 3, and displays the conversion candidates for which confirmation of conversion has been instructed. It controls various processes such as writing into the text memory 4, displaying on the display section 3, printing on the printing section 5, and storing in the external memory 6.

In addition to the kana key 2a for inputting kana characters, the key human power unit 2 includes a conversion key 2b for instructing to convert the input kana character string into a kana-kanji mixed character string, and an instructing to convert it to another conversion candidate. Next candidate key 2C1 Various function keys such as a confirmation key 2d for confirming a conversion candidate, a reconversion key 2e for instructing to reconvert a once confirmed conversion candidate to another conversion candidate, and a cursor key 2f for specifying a position, etc. have.

The input buffer 7 temporarily stores the kana character string input from the key input section 2.

Kanji characters and their pronunciations are preset in the conversion dictionary memory 8 in correspondence with each other, and the determined conversion candidates are stored in a clause unit in the learning memory 9 in association with the corresponding kana character strings.

The conversion unit 10, under the control of the CPUI, has a learning memory 9,
Referring to the conversion dictionary memory 8, the kana character string stored in the input buffer 7 is converted into kana-kanji characters to generate a kana-kanji mixed character string.

The kana-kanji mixed character string generated by this conversion unit 10 is
Under CPU control, candidate buffer 1 is selected as a conversion candidate.
1 and displayed on the display unit 3.

Note that the learning memory 9 is, as described above, the large power buffer 7.
In addition to being used when converting a kana character string stored in It is used to accurately extract phrases that include the specified character. In performing this cutting, CP
The UI utilizes the extraction buffer 12, search pointer S, and character count register N.

Next, the operation of the embodiment will be explained with reference to FIGS. 3 to 6.

First, the kana-kanji conversion process will be explained with reference to FIGS. 3 and 4.

When the conversion key 2b is operated, the conversion unit 10
Under the control of the learning memory 9, the CPU 1 performs conversion analysis to convert the kana character string in the input buffer into kana-kanji (step S1 in FIG. 3). It is determined whether conversion is possible (step S2). If the result is that conversion is possible, the conversion candidate (kana-kanji mixed character string) obtained by the kana-kanji conversion is stored in the candidate buffer 11. S3)
, is displayed on the display unit 3 (step S4), and the operator judges the displayed content and performs key operations to instruct various functions (step S5). In this case, the next candidate key 2C is operated. If so, the conversion unit 10 refers to the conversion dictionary memory 8 and performs conversion analysis to convert the kana character string in the input buffer into kana-kanji (step S6).
Note that if it is determined in step S2 that conversion is impossible, the process immediately proceeds to step S6. In other words, if the first conversion candidate cannot be generated based on the learning memory 9, the first conversion candidate cannot be generated based on the conversion dictionary memory 8. to generate a first conversion candidate.

When the confirmation key 2d is operated in step S5,
The CPU 1 stores the conversion candidates in the candidate buffer 11 in the text memory 4, and determines the conversion contents (step S7).
). Then, the confirmed conversion candidates and their pronunciations are associated in units of clauses and stored in the learning memory 9 (step S8), input buffer, and candidate buffer 11.
The contents of are cleared (step S9), and the process ends.

Next, the kana-kanji conversion process shown in FIG. 3 will be explained in detail with reference to FIG. 4.

Currently, the text memory 4 stores the document data "I understand that you are on a long voyage," as shown in FIG. 4(a), and the learning memory 9 stores the document data.

The data in the part ■ in Figure 4(b) is stored, and the input buffer 7 contains the pseudonym ``This year's heat is very hot'' as shown in Figure 4(C). Assume that the conversion key 2b is operated while the character string is temporarily stored.

In this case, since the learning memory 9 does not have a pronunciation corresponding to the kana character string in the input buffer, conversion candidates are generated based on the conversion dictionary memory 8. That is,
A conversion candidate group of four clauses, ``This year's thickness is very above,'' as shown in FIG. 4(d), is stored in the candidate buffer 11 and displayed on the display unit 3. At this time, in the candidate buffer 11, a segment break mark (not shown) is inserted.

Of this conversion candidate group, ``This year's thickness is very thick,'' I noticed that the phrase ``That's it, J'' was incorrectly converted, so I specified the phrase ``That's it,'' with the cursor, and next When the candidate key 2c is operated (corresponding to step S5), as shown in FIG. and displayed on the display unit 3. When a confirmation process is instructed for the conversion candidate group shown in FIG. 4(e) using the confirmation key 2d, the contents of FIG. 4(e) are stored in the text memory 4 as shown in FIG. 4(f). Newly stored. Also,
In the learning memory 9, the conversion candidate group shown in FIG. 4(e) is newly registered for each phrase in association with its pronunciation.
(See ■ in figure (b)).

Next, the reconversion process will be explained with reference to FIGS. 5 and 6.

When performing re-conversion, there is no need to specify the character string to be re-converted with consideration given to the break between clauses; it is sufficient to arbitrarily specify one desired character in the text memory 4; If you use the cursor key 2f to specify any character in a character string that has an error in the conversion result on the document data that has been converted, and operate the reconversion key 2e, a predetermined number of characters centered around the character specified by the cursor will be displayed. Extract the character string from the text memory 4 and store it in the extraction buffer 12 (
(Step 551 in FIG. 5), at this time, in this embodiment, the capacity of the extraction buffer 12 is for 13 characters, so
Thirteen characters centered around the character specified by the cursor are extracted and stored in the extraction buffer 12.
As shown in Figure 6(a), when the position of ``Sa'' in the phrase ``Thickness'' is specified with the cursor, as shown in Figure 6(b), ``Sa'' is the center point. The 13 characters ``However, this year's thickness is very abnormal'' are extracted and stored in the extraction buffer 12.

Then, the digit position ``'7'' (see Figure 6<b)) of the specified character on the extraction buffer 12 is initialized to the search pointer S (step 552), and the character number register N
is initially set to "7" (step 353).

Next, the learning memory 9 is searched for a character string of the number of characters indicated by the number of characters register N after the digit position indicated by the search pointer S in the extraction buffer 12 (step 5
54) Then, it is determined whether the corresponding character string exists or not (step 555). If it does not exist, the contents of the character number register N are decremented by '1'' (step 856), and the contents of the character number register N are "0" (step 557), and as a result, "O".
If not, the process returns to step S54. on the other hand,
When the content of the character count register N reaches °゛0゛, as can be inferred from N=7, N=6...N=O shown in Figure 6(b), the current search point S This means that the character string after the indicated digit position does not match the content of the learning memory 9 no matter how many characters are compared, so the content of the search pointer S is decremented by "1" (step 558), and the search is performed. It is determined whether a character exists at the digit position indicated by pointer S (step 55).
9) As a result, if the search pointer S exists, return to step 953 and repeat the same process.
When the content of is ``6'' and the content of the character count register N is ``7'', as is clear from the part where N = 7 shown in Figure 6(C), ``the thickness is very different'' is stored in the learning memory. Search on 9.

When it is determined in step S55 that the character string specified by the search pointer S and the character count register N exists in the learning memory 9, the corresponding pronunciation string is read from the learning memory 9 and stored in the input buffer 7. (Step 560). In the above example, as shown in FIG. When the contents of register N become 3゛ and the character string defined by these becomes "Thickness wa", the pronunciation string "Atsusaha" corresponding to "Thickness wa" is stored in the input buffer 7. (See FIG. 6(d)).

Next, referring to the conversion dictionary memory 8, conversion analysis is performed to convert the kana character string in the input buffer into kana-kanji (step 561), and the conversion candidates obtained by the kana-kanji conversion are stored in the candidate buffer 11. Step 562
) is displayed on the display unit 3 (step 963).The operator judges the displayed content and performs key operations to instruct various functions (step 564).In this case, the next candidate key 2C is operated. If so, step S61
Convert the kana character string in the input buffer to another conversion candidate by returning to , is converted into the next candidate "Natsu wa" shown in FIG. 6(f).

In step S64, when the confirmation key 2d is operated, the number of characters of the previously confirmed conversion candidate and the candidate buffer 1
It is determined whether the number of characters of the conversion candidate to be confirmed this time stored in 1 is the same (step 565).
As a result, if the number of characters is different, the entire document data after the previously confirmed conversion candidate in the text memory 4 is shifted by the difference in the number of characters (step 866). If the number of characters in the conversion candidates that have been confirmed is greater than the number of characters in the conversion candidates to be confirmed this time, by shifting the number of characters to the left by the number of characters in the conversion candidate to be confirmed this time, delete the character string that has the largest number of characters among the conversion candidates that were confirmed last time. do. Also, if the number of characters in the previously confirmed conversion candidate is less than the number of characters in the currently confirmed conversion candidate, the number of characters in the currently confirmed conversion candidate is shifted to the right by that small amount. Secure the area.

Then, the conversion candidates in the candidate buffer 11 are confirmed by overwriting the first position of the previously confirmed conversion candidates in the text memory 4 (step S67, see FIG. 6(g)).
.

For example, as shown in FIG.
In the case where the current confirmation instruction is given when the conversion candidate in 1 is "Natsu wa", in step S65, the number of characters of the previously confirmed conversion candidate and the number of characters of the conversion candidate to be confirmed this time are determined. If it is determined that they match, step S66
The confirmation process in step S67 is performed without performing the shift process in step S67.

Next, on the learning memory 9, by shifting the part after the previously determined conversion candidate "Thickness wa" and the corresponding pronunciation "Atsusawa" forward by one phrase, Delete the pronunciation "Atsusaha" that corresponds to the above "Thickness" and use the conversion candidate "Natsusaha" in the candidate buffer 11.
The corresponding pronunciation "Atsusaha" is newly registered in the empty area at the forefront (step 968, see FIG. 6(h)).

That is, the conversion candidates and their pronunciations are stored in the learning memory 9 in the order in which they are determined.

Then, the contents of input buffer 7 and candidate buffer 11 are cleared (step 569), and the process ends.

In this way, by simply specifying one character in the confirmed document data, the clause containing the specified character is accurately extracted and reconverted.

[Effect of the invention] According to this invention, any one on the confirmed document data
Even if a character is specified, the phrase containing this character can be accurately extracted, and even if the delimiter of the phrase is not known, it is possible to perform accurate reconversion.

[Brief explanation of drawings]

Fig. 1 is a functional block diagram of the present invention, Fig. 2 is a block diagram of an embodiment, Fig. 3 is a flowchart showing the kana-kanji conversion process, and Fig. 4 specifically explains the kana-kanji conversion process according to Fig. 3. FIG. 5 is a flowchart showing the re-conversion process, and FIG. 6 is a diagram specifically explaining the re-conversion process according to FIG. 5. DESCRIPTION OF SYMBOLS 1...CPU, 2...Key human power department, 2b...Conversion key, 2d...Confirmation key, 2e...Reconversion key, 2
f--Cursor key, 4--Text memory, 7-
・・Incover Nofa, 8・Conversion dictionary memory, 9・・
- Learning memory, 10... Conversion unit, 12... Extraction buffer, S... Search pointer, N... Character number register. Patent applicant Casio Computer Co., Ltd. Figure 2 Figure 3

Claims (1)

[Claims] Based on the conversion dictionary storage means that stores kanji and their pronunciations in correspondence, the input kana character string is converted into a kana-kanji mixed character string in units of phrases, and the kana of each phrase is converted into a kana-kanji mixed character string. After determining the kanji mixed character string, the kana-kanji conversion device that writes each kana-kanji mixed character string as document data into the document storage means converts the kana-kanji mixed character string of each clause determined above and the corresponding kana character string. A fixed clause storage means that stores associated phrases, and when an arbitrary character position on document data in the document storage means is specified under a reconversion command, kana-kanji mixed characters before and after the specified character position. a search means for searching for a character string that matches the string by comparing it with each kana-kanji mixed character string stored in the fixed phrase storage means; A kana-kanji conversion device comprising: reconversion means for reading a kana character string from the determined phrase storage means and reconverting the kana character string based on the conversion dictionary storage means.