JPS5880771A - Kana (japanese syllabary)-chinese character conversion processor - Google Patents

Abstract

PURPOSE:To facilitate an easy conversion into a desired sentence including Chinese characters, by selecting and deciding successively the candidate words of the suffix and plural independent words and their accessory words which can form a clause within a range where the relation of connection is grammatical among those above-mentioned words. CONSTITUTION:A connection inspecting circuit 6 inspects the connection between independent words and accessory words and delivers an inspection output E. A character train coincidence detecting circuit 7 checks whether or not the suffix, independent words and suffix candidate in the KANA letter train exist in the independent, suffix and prefix tables 3-5. At the same time, a control circuit 2 feeds the residual character trains which function as the conjugation D of parts of speech and the candidates of accessory words to the circuit 6 by the output E of the circuit 6 and an indication given from a keyboard 1. Thus the independent words are selected or fixed independently of each other while abbreviating properly the suffix, independent words and prefix candidate of the KANA letter train and at the same time satisfying the relation of grammatical connection among these words.

Description

Detailed Description of the Invention The present invention provides a kana-kanji conversion processing device that converts and outputs a kana character string input from a kana character input unit such as a kana keyboard into a sentence containing kana-kanji. The present invention relates to a kana-kanji conversion processing device configured to sequentially select and determine a plurality of candidate words for independent words and adjunct words within a range in which their connection relationships comply with the grammar.

In general, it is well known that the typical form that makes up a clause is (prefix) + (independent word) n + (suffix) + (adjunct word).Here, () is the prefix mentioned above, the independent word , suffixes, and adjuncts exist or do not exist as part of the clause; ()n must exist at least once;
(') m indicates that there is a plurality of each or there is no presence.

Now, the basic conversion procedure for converting a kana character string input from a kana keyboard into a kanji-kana mixed sentence in a conventional kana-kanji conversion processing device will be explained with reference to the flowchart shown in FIG.

In step S1, an independent word that can be a candidate is searched in the independent word table according to the longest match principle for the input kana character string, and if this independent word is found, in step S2, an adjunct word is searched from the remaining character strings. , and check the validity of the connection with the independent word extracted earlier
Check in step 3. If this validity pillar is good, each independent word and attached word candidate found here is immediately output as a clause word candidate as a result of kana-kanji conversion.

On the other hand, if an attached word that can be a candidate is found but the connection test with the previously discovered independent word fails, further step S4
The above step s2. It is determined whether the process of s3 has ended.

If this determination is negative, the process returns to step S2 to further search for other candidates that can be used as adjunct words, and performs the same process. Also, if an appropriate adjunct cannot be found from the remaining character strings for the previously discovered independent word, use another independent word.

For example, if there are no candidates for a group of homophones that make up the same kana heading, the target independent word kana character length is further shortened and other independent words are searched, and as mentioned above, the independent word and independent word are searched for. 1. Then, in step S5, if all independent word candidates fail in testing with adjunct words, the next step is to extract independent words in step S6,
In step S7, the suffix table group is searched to detect whether a suffix exists in the head part of the kana character string excluding independent word candidates, that is, the above-mentioned adjunct word candidate. However, if a suffix is found, it is decomposed into a suffix and an adjunct, and in step S8, an adjunct test is performed on the remaining kana character string. If the connection test fails here, it is determined in step S9 whether or not all suffixes and adjuncts have been checked. If not, the process returns to step S7; if not, the process proceeds to step S10; In this case, it is determined whether all independent words have been tested. If the result of this determination is negative, the process returns to step S6; otherwise, the process proceeds to step 513, in which it is determined whether or not there is a prefix for the beginning part of the input character string; For one prefix candidate, the remaining part of the input string is tested for the connection of the independent linguist adjunct and the independent linguist suffix + adjunct, and if all tests fail, it searches for other prefix candidates. , perform the above series of processes on this prefix candidate.

In addition, in FIG. 1, 811 is the step 51 to S5 described above.
This step consists of extracting independent words and attached words from a kana character string and testing the connection between the two. Also, 5
12 is a step consisting of the above-mentioned steps 56 to S10, which decomposes the kana character string into independent linguist suffix adjunct words and performs a connection test on the adjunct words; 813 extracts prefixes and words from the ephemeral character string; This is a step in which the remaining portion is subjected to the processing of both steps Sll and 312 described above.

814 is a step of outputting that the kana-kanji conversion has been successful, and S15 is a step of outputting that the kana-kanji conversion has failed.

In addition, the failure of the test here refers to the connection relationship between words 2, for example, in addition to cases in which word succession cannot be established grammatically in Japanese, even if the connection relationship is established grammatically. For example, it also means a case where the word structure intended by the operator cannot be obtained. In such cases, a method is usually adopted in which the operator instructs the system that the test has failed by giving a command to proceed to the next test.

Conventional kana-kanji conversion processing devices perform the above processing to obtain kanji-kana mixed sentences, but in Japanese there are homophones with different meanings for the same kana heading, and these are independent words. The same applies to prefixes and suffixes in addition to words. Now, just the combination of colloquial language + suffix is P.

There are as many combinations as XPIxP2, and there is a drawback that it takes a lot of time and effort to find the kanji/kana mixed sentence that the operator intends.

The present invention has been made in order to eliminate the drawbacks of the conventional methods as described above. Select each independently in a state that satisfies the connection relationship,
It is an object of the present invention to provide a kana-kanji conversion processing device that can significantly reduce the time required for kana-kanji conversion by providing a partial confirmation input means that can be confirmed.

An embodiment of the present invention will be described below with reference to the drawings.

FIG. 2 shows a kana-kanji conversion processing device according to an embodiment of the present invention. In the figure, (1) is a keyboard as an ephemeral character string input section, which is used as an operator to enter partial confirmation and change the above candidates. It has a candidate change input key as a candidate change input means for inputting a candidate change. Also (3
) is an independent word table, which has a kana-to-kanji conversion pair consisting of a kana header section and a kanji header section, as well as information such as part of speech and conjugation form for each independent word. +41 +
Reference numeral 51 denotes a suffix table and a prefix table, which have only kana and kanji conversion pairs consisting of a kana header part and a kanji header part. In addition, each table (3) to (5)
Tomokana headings are given codes in ascending order of the 50 syllabary, and the tables are also arranged in the order of the 50 syllabary. (
6) is a connection test circuit that tests the connection between the above independent word and adjunct word and outputs the test output E, and (7) is a character string match detection circuit, which detects the suffix and independent □ in the above kana character string. It is detected whether word and prefix candidates exist in the above tables (3) to (5). In addition, (2) uses the test output E of the connection test circuit (6) to appropriately shorten the suffix, independent word, and prefix candidates of the kana character string, and to convert the above independent words, etc., while satisfying the grammatical connection relationship. This is a control circuit that selects or confirms independently.

FIG. 3 is a flowchart showing an outline of the processing procedure of the apparatus of the above embodiment, and in the figure, 821 is a step for initializing control statuses ST and i, S22
823 is a step of determining whether the control status ST is 2 or 4, S24 is a step of extracting a suffix, and S25 is a test of connection between an independent word and an attached word. Step, S26, step of outputting the result of kana-kanji conversion, S27. S28゜S
29 is a step of determining whether the above conversion result should be changed, partially confirmed, or fully confirmed according to the instructions of the operator, and 529a is a step of determining whether the conversion result should be changed, partially confirmed, or fully confirmed, respectively, and 529a is a step of determining whether the Kana-Kanji conversion is successful after determining the complete confirmation in step 829. S30 is a step of increasing the value of control status i by 1, 531 is a step of determining whether control status ST is 4 or more, and 531a is a failure of kana-kanji conversion when "YES" in step 531. 832 is a step of increasing the value of the control status ST by 1, S
33 is a step of determining whether the control status ST is 3 or 4, S34 is a step of extracting a prefix, and S35 is a step of forcibly setting the value of control status i to 1.

FIG. 4 shows an example of the configuration of the character string match detection circuit in the apparatus of the above embodiment. In the figure, (1 old control circuit (2)
A buffer register into which a character string is input (12
is an address register in which the address of each table (3) to (5) is set, (131 is each table (3) to (5) above)
This is a matching circuit that determines whether the contents of (5) are included in the contents of the buffer register fill, and this matches the contents of the buffer register +Ill in the table (31 (4) or (5)). When there is something, the output line (
13a), outputs an output S indicating successful detection, and if there is no output, outputs an output that increments the address stored in the address register (121 by 1) via the output line (13b). When an independent word, suffix, or prefix is removed, and the removal fails, that is, when the independent word, etc. is longer than the input kana character string, the output line (1
4a) is a removal circuit that outputs an output F indicating a detection failure.

Next, the operation will be explained using FIGS. 2 to 4.

First, a kana character string input from the keyboard (1) is stored in an input buffer register (not shown) of the control circuit (2). Next, when the kana-kanji conversion trie is given, in step S21, the control status ST and i are set to ST = l and i = 2, respectively, and
The contents of the input buffer register are processed by the string match detection circuit (7
), and only TSL is significant as a table selection signal. Now, an example of the character string match detection circuit (7) will be explained in the fourth example.
As shown in the figure, in this circuit (7), the character string sent from the control circuit (2) is stored in the second buffer register (11).
is temporarily stored. Next, the search start address of the search table (in this example, to simplify the explanation, the search is assumed to start from the beginning of the table) is set in the address register (1
21, and its output TA is the search specification table,
That is, in this case, it is sent to the independent word table (3). Then, the output information T I) read from this independent word table (3) is the kana heading part (hereinafter referred to as B).
is sent to the matching circuit (13), and this matching circuit (]3
), the contents of the second buffer register (11) (
It is checked whether or not character string B exists in character string A in the buffer register (11) using the first character of character string A (hereinafter referred to as A) as a reference. Then, when the character string B does not exist in the character string A, the matching circuit (13) outputs the output line (13b
), the contents of the address register (12) are incremented by +1, the next independent word in the independent word table (3) is read out, and the same process is repeated one after another.

Now, when the contents of buffer register (11) do not exist in the independent word table (3), that is, buffer register 01
), there is no kana heading part of the output information TD, that is, the character string B, and the difference A-B between the content of the buffer register (11), that is, the character string A, and the above character string B is A-B (
When Q, the output F indicating detection failure is connected to the output line (14
a) to the control circuit (2).

On the other hand, when the character string B is found in the buffer register (11), it is determined that the detection is successful, the table search is temporarily stopped, and an output S indicating that the detection is successful is sent along with the output information TD through the output line (13a). and sends it to the control circuit (2). That is, this is the details of the process in step S22. In this way, if the independent words are successfully extracted, step S23
Since 5T=1 now, a path passing step 824 is selected, and then in step 525, the connection test between the independent word and the attached word is performed in the connection test circuit (6).

For this connection test, the parts of speech and inflected forms of the read independent word information, the remaining character string (hereinafter referred to as C) obtained by removing character string B from character string A, and the connection test table are used. . This connection test table, for example, arranges parts of speech and conjugations on the horizontal axis and adjuncts on the vertical axis, and when the two are connected grammatically or idiomatically, 1. If there is no character string, the character string to be checked is determined according to the principle of O.

Then, when the connection test is successful in step 825 and the kana-kanji conversion result is output in step S26, the operator looks at the display (8) and if the result is what the operator desires, presses the confirmation key. If some of the candidates are desired, the user presses the partial confirmation key, and if the candidates are not desired and the candidate is to be changed, the candidate change input key is pressed. Then, this device performs step 8.
When the confirmation type key of 29 is pressed, step 52
9a is reached, and a message indicating that the kana-kanji conversion was successful is output. Further, when the partially determined key is pressed in step 828, the extracted independent word is determined to be correct, the control stick i=3 is set in step 530, steps S23 and S24 are passed, and step 82 is performed.
In step 5, a connection test is performed between the independent words determined to be overwhelming and the attached words that are the remaining parts thereof. Furthermore, when the candidate change entry key in step S27 is pressed, the process returns to step S22 to re-detect the independent word, add 1 to the address register (121), search for another independent word,
Similarly, in step S25, the remaining portion, that is, the attached word, is tested for connection between the independent word and the attached word. If the connection test fails for all the independent words, the process goes to step S3] and is set to 5T-2 in step 832, and the process moves to the connection test for the independent word suffix adjunct words. 5T-2
When this happens, the process immediately proceeds from step S33 to step 82.
2, after extracting independent words from the entire input character string in step 322, it is decided in step S23 to execute step S24, and in this step S24, it is determined whether there is a character string that can be a suffix in the remaining part. suffix table (4) (in this case, the table selection signal is used to search for another independent word when no suffix candidate is extracted in the above step 824 of rS2. In step S24, a suffix candidate is extracted) Then, in step S25, the residual part (attached word) is tested for connection with an independent word.If the extracted independent word or suffix fails in the connection test with an attached word, the test is performed in step S24. Search for other suffix candidates in If the connection test between the independent word extracted as a word candidate and an adjunct word fails, another independent word candidate is searched for.If the test of all independent word candidates and suffix candidates fails, 5T= 3, execution of step S34 is selected in step S33, and the process moves to the prefix+independent word"-adjunct word verification process.Now, if the independent word and suffix test is successful in the above,
In step S26, the result is output to the display (8).

Here, the operator judges whether or not the desired kana-kanji conversion has been performed while looking at the display (8).

Then, when the operator operates the candidate change input key in step 827, the system returns to step S22 and starts searching for another independent word as if the connection test had failed. Also, if there is a partial confirmation input key operation in step 828, it is determined that the independent word extracted earlier is correct, and i=3 is set in step 530, and the process returns to step S24, and the suffix is selected from the input remaining character string. Let's move on to the process of extracting words. Then, after entering this mode, in step 825, does the connection test between the remaining part (adjunct word) after removing the suffix and the previously determined independent word fail?

Or even if it is successful, the operator instructs to search for another suffix in step S27 (using the candidate change input key).
If there is, the system searches for other suffix candidates and performs the same process as above. By the way, when an independent word is partially confirmed or the connection test between all suffixes, adjuncts, and independent words in the input residual string fails, the independent word that was partially confirmed first is treated as correct. (However, in this case, the independent word that was partially determined earlier is also treated as an error and s −r
= 3. ) When ST=3, prefix extraction is started in step S34 to determine whether there is a prefix at the beginning of the input character string.

Here, if the prefix does not exist, it is assumed that the kana-kanji conversion has failed and the kana-kanji conversion is terminated. On the other hand, if a prefix is found in step S34, the input kana character string from which the prefix has been removed is tested for connection of independent linguists' adjuncts in step S25. Here, if the independent word and adjunct word connection test is successful for the candidate prefix, the result is outputted to the display (8) in step 526. Here, when the candidate change input key in step S27 is pressed, another prefix candidate is found and the same process is performed on the remaining part. On the other hand, when the partial confirmation key in step 828 is pressed, the previously extracted prefix is determined to be correct, and the remaining portion is subjected to the same processing as ST:1.

Finally, if the connection test fails for all prefix and independent word candidates, the process moves to s - r = 4. This is an examination of the composition of prefix masters, independent language masters, suffixes + adjuncts. That is, first, the prefix is extracted in step S34, and then in step S22. In step S24, independent words and suffixes are extracted, and in step S25, a connection test of independent language expert adjunct words is performed. If the connection test fails here, it means that the kana-kanji conversion has failed.On the other hand, if the connection test is good, the conversion result is output in step S26, and if the confirmation key is pressed in step S29. The stupid kanji conversion was successful, and the candidate change input key 2 in step 827 or 828
If the partial confirmation key is pressed, step S3 is executed.
4 and step S22 to re-extract the prefix and independent word.

If the connection test fails as a result of all of the above processing, it is assumed that the kana-kanji conversion has failed, and in step 531a, this is prevented and the entire input character string is output as is. Also, if there is a confirmation input key operation in step S29 during the process, in step 529a, the kana-kanji conversion is successful and the prefix, independent word, suffix, and attached word are output as a mixed sentence of kanji and kana. .

In addition, the above-mentioned switching of the table selection signal TS, setting of the remaining part after removing the independent word part from the input character string to the buffer register (11) in the character string match detection circuit (7), and extraction of the suffix were also carried out. The setting of the remaining part of the character string to the connection verification circuit (6), etc. are all controlled by the control circuit (2).

As described above, according to the present invention, when each candidate of a prefix, an independent word, and a suffix appears (extracts), a partially definite formula key is provided for each candidate, and this partially definite formula key is inserted into the input kana character string. A prefix that can exist multiple times.

Since the independent words or suffixes intended by the operator are partially determined, the input character string is divided into two groups, each with p, p, and suffix candidates, respectively. Even if p,..., the clause composed of them is almost P. +p, 10p2+, . . . or less can be easily converted into a desired sentence containing kana and kanji.

[Brief explanation of drawings]

FIG. 1 is a flowchart showing the processing performed by a conventional device, FIG. 2 is a block diagram showing a kana-kanji conversion processing device according to an embodiment of the present invention, and FIG. 3 is a flowchart showing the processing performed by the device of the above embodiment. FIG. 4 is a block diagram showing an example of the circuit configuration of the character string match detection circuit of the above embodiment device. (1)...Kana character input section, partial confirmation input means, candidate change input means (keyboard), +31...Independent word table, (4)...Suffix table, (5)...Prefix Table, (7)...Character string match detection circuit, (6
)... Connection verification circuit, (2)... Control circuit, (8)
···display. Note that the same reference numerals in the figures indicate the same or equivalent parts. Agent Makoto Kuzuno - Figure 3 Figure 2 Figure 4

Claims (1)

[Claims] (1) A kana character input section for inputting a kana character string,
An independent word, suffix, and prefix table that stores kana-kanji conversion pairs related to independent words, suffixes, and prefixes, and candidates for independent words, suffixes, and prefixes in the above input Kana character string are stored in the above table. a character string match detection circuit that reads out and outputs the candidate from the table when it exists in each of the above characters; a partial confirmation input means for giving an instruction to the character string match detection circuit displayed on a display device to be described later; and the display device. a connection testing circuit that tests whether or not the connection relationship between adjunct word candidates i excluding independent words, suffixes, and prefixes is grammatically correct, and the above-mentioned part The above-mentioned independent words 1. It is equipped with a control circuit that independently determines the suffix or prefix candidates while changing them to other candidates as necessary, and a display that displays the output of this control circuit, which is a sentence containing kana and kanji. Features: Kana-Kanji conversion processing device.