JPH0628057B2 - Character processor - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial application] The present invention relates to a character processing device for inputting a kanji / kana mixed sentence by kana / kanji conversion.

[Prior Art] Currently, a character processing device such as a Japanese word processor generally inputs kanji-kana mixed sentences using kana-kanji conversion.

Kana-Kanji conversion is to convert an input reading string into Kanji by referring to a dictionary. Part-of-speech information such as nouns, sahen nouns, adverbs, adjectives, and adjective verbs is described for each word in the dictionary, and Kana-Kanji conversion analyzes possible reading strings to create possible phrase candidates and combines them. The conversion candidates are determined and presented in order of likelihood. The operator selects a desired candidate from the presented conversion candidates.

For example, for the reading line "I didn't know what to say", "stomach", "committee", "committee", "to committee", "mark", "mark"
"Flame""Division""Meeting""Meeting""Stomach""Stomach""Stomach""Two""Two""Leaf""Measured""Planned""Consulted""Bought The phrase candidates such as “” are created, and “commit to committee”, “plan for committee”, “consult committee” are combined and output as conversion candidates in this order. Of the “measure”, “inquire”, and “consult”, “measure” and “inquire” are generally more frequent than “inquire”, so “inquiry to the committee” is not converted as the first candidate.

Therefore, in order to increase the conversion rate, a method called example conversion has been proposed. The example conversion is a method in which a pair of co-occurring words (examples) is registered in the example dictionary in advance, the example dictionary is referred to during conversion, and the first candidate is changed according to the example in the example dictionary. For example, the pattern “I consult / commit to the committee” is stored in the example dictionary, and “I consult the committee” is the first for the input “I was not good at it”.
Make it converted as a candidate.

Although the above example is an example of an individual example in which the relationship between words is described, a semantic example in which the relationship with a semantic classification is described is also proposed. For example, in the context of “to ni,” when “to” represents a place,
The notation is plausible, and when "~" represents a human, the notation "said" seems plausible. In order to perform this conversion correctly, the meaning examples of "to / to say" to "human" / to / to "place" are registered in advance in the example dictionary. Note that “human”, “place”, etc. are semantic classifications, and all words in the dictionary describe “human”, “place”, or not. With this mechanism, it is possible to correctly convert "I told the mayor,""I went to the city hall," etc.

[Problems to be Solved by the Invention] However, in the conventional example conversion technique, when a plurality of examples are applicable, that is, when the examples conflict, which example should be prioritized is specified. There wasn't. Even if there was a slight guideline to give priority to individual examples when there was a conflict between individual examples and meaning examples, it was not clear which example should be given priority in other situations. Therefore, since the first candidate was appropriately decided by adopting an appropriate example, the conversion rate in a delicate situation was not good.

[Means (and Action) for Solving the Problems] In order to solve the above problems, the character processing device of the present invention is
An input means for inputting a kana character string, a word dictionary means that stores a word reading, a notation and a meaning classification in association with each other, a word notation, and a notation or a meaning classification of a word used in combination with the word And a candidate creating means for creating a notation candidate in which the kana character string input by the input means is read by referring to the word dictionary means. Regarding the candidates of the notation for the same reading created by the creating means, in the example stored in the example dictionary means, the corresponding candidate is given priority along with the word to be used together, and when there are a plurality of applicable candidates, the candidate and the By prioritizing candidates that correspond to examples in which the distance to the words used together is close in the sentence,
Display priority determining means for determining the display priority of the notation candidates, and candidate display for displaying the notation candidates created by the candidate creating means based on the display priority determined by the display priority determining means. And means.

EXAMPLES The present invention will be described in detail below with reference to the drawings.

FIG. 1 is an example of the overall configuration of the present invention.

In the illustrated configuration, the CPU is a microprocessor, performs arithmetic operations for character processing, logical judgments, etc., and has an address bus AB, a control bus CB, and a data bus D.
Via B, each component connected to those buses is controlled.

The address bus AB transfers an address signal indicating a component to be controlled by the microprocessor CPU. Control bus CB is a microprocessor CPU
The control signals of the respective components to be controlled by are transferred and applied. The data bus DB transfers data between the constituent devices.

Next, the ROM is a fixed read-only memory.
Microprocessor CP described later with reference to FIGS. 0 to 15
The control procedure by U is stored.

The RAM is a writable random access memory having a structure of 1 word 16 bits, and is used for temporary storage of various data from each constituent element. IBUF is an input buffer that stores key data input by a key input.
UF is an output buffer for temporarily storing the result of Kana-Kanji conversion. DIC is a dictionary for Kana-Kanji conversion. YDIC is an example dictionary that describes a co-occurrence relationship between words that have not been classified. PRTBL is a semantic classification priority table describing a priority relationship between semantic classifications.
BCTBL is a phrase candidate table that stores phrase candidates that are in the process of conversion.

KB is a keyboard, which is a character / symbol input key such as an alphabet key, a hiragana key, or a katakana key, and
Various function keys for instructing various functions for the character processing device such as a conversion key are provided.

DISK is an external storage unit for storing document data, and stores a document created in the text buffer, and the stored document is called when necessary by a keyboard instruction.

CR is a cursor register. The CPU can read and write the contents of the cursor register. The CRT controller CRTC described later displays a cursor at a position on the display device CRT corresponding to the address stored here.

DBUF is a display buffer memory that stores data to be displayed.

CRTC is cursor register CR and buffer DBUF
It plays the role of displaying the contents stored in the display CRT.

The CRT is a display device using a cathode ray tube or the like, and the display pattern of the dot configuration and the display of the cursor on the display device CRT are controlled by the CRT controller.

Further, CG is a character generator, which stores patterns of characters and symbols to be displayed on the display device CRT.

The character processing device of the present invention comprising the above-described components operates in response to various inputs from the keyboard KB. When an input from the keyboard KB is supplied, first, an interrupt signal is sent to the microprocessor CP.
The control signal is sent to U, the microprocessor CPU reads out various control signals stored in the ROM, and various controls are performed in accordance with these control signals.

FIG. 2 is a diagram showing a conversion example by the device of the present invention. In the figure, "" means that the conversion key is pressed there.

If you enter "Iinkai Ikari", it will be converted to "Consult the Committee" as the first candidate. This is because the example dictionary "/ consult to the committee" is registered. If there is no such example conversion mechanism, it will be "measured"
For, there are homophones such as "measure,""inspire," and "attempt," so they are not always converted correctly.

The following example is for the case of inputting “measure kaiketsu”.
Similar to the first example, since the example “solve / seek” is registered in the example dictionary, it is correctly converted to “solve”.

The third example is a conversion in the case of inputting that "you have asked to check your score". In this case, the two examples of “consulting with the committee” and “solving” can be applied, and each of the “measure” parts requires contradictory conversion. In other words, the conversion result of "measure" changes depending on which example is believed. According to the present invention, in such a case, conversion is performed on the basis of the example in which the distance between phrases is shorter. That is, the example of “to consult / consult” is applied across one phrase of “solve”, but the example of “to solve / intend” can be applied between two consecutive phrases. Therefore, the example of “solving / aiming” having a shorter distance is adopted, and it is converted to “requesting the committee to solve”.

FIG. 3 shows an input buffer IBUF and an output buffer OBUF.
It is a figure showing the composition of.

Both IBUF and OBUF have the same configuration. The first 2 bytes are the size information of each buffer, and the numerical value obtained by doubling the value obtained by subtracting 1 from the number of characters stored in the buffer is entered. The "" at the end of the input buffer means that the conversion key was pressed there. Each character consists of 2 bytes per character and is stored in JIS X 0208 code or the like.

FIG. 4 is a diagram showing the structure of the dictionary DIC. "reading"
It is composed of fields of "notation", "part of speech", "word likelihood", and "semantic classification".

"Reading" means reading a word, "writing" means writing a word,
The "part of speech" stores the part of speech of the word.

In the "word likelihood", information indicating the likelihood of the word itself such as frequency information is stored as a value of 1 to 5. The likelihood value 5 means that it is most likely, and the smaller the value, the more suspicious. A likelihood value of 0 does not exist as a word likelihood value because it means that it cannot be considered at all.

In the "semantic classification", the semantic classification of the word is generally described as "organization", "animal", "food", etc. However, semantic classification is described only when the word is a noun.

FIG. 5 is a diagram showing the structure of the example dictionary YDIC.
It is composed of fields of "first word", "second word" and "particle".

A pair of co-occurring words is described in “first word” and “second word”. If it is not a specific example but a semantic example, a semantic classification is described instead of a word.

The "particle" describes a particle that connects a pair of words.

That is, the example shown in the figure is "to consult / consult"
"Solving / aiming""Animal" / keeping "" Food "/
To buy ”.

FIG. 6 is a diagram showing the structure of the meaning classification priority, which describes the priority relationship between the meaning classifications. It is composed of fields of "semantic classification" and "priority".

The "semantic classification" stores a semantic classification code, and the "priority" stores a priority for the semantic classification. As the priority, information indicating the likelihood of the semantic classification is stored as a value of 1 to 5. The likelihood value 5 means that it is most likely, and the smaller the value, the more suspicious.

FIG. 7 is a diagram showing the concept of the phrase candidate table BCTBL. The phrase candidate table is a binary phrase representing possible phrase candidates as a result of analysis of input reading. In the figure, a horizontal line means a child pointer, and a vertical line means a younger brother pointer. The younger brother pointer links other bunsetsu candidates (usually shorter ones) starting from a certain reading position, and the child pointer links bunsetsu candidates that follow the bunsetsu.

At the beginning of the input reading string "I asked you to make a call to a member", there are interpretations such as "in the committee", "committee", "committee", etc. It is linked.

The phrase following "to the committee" may be "solve",
It is linked by a child pointer.

When such a clause candidate table is created, "I asked the committee to try to solve it,""I asked the committee to consult me for the solution,""I asked the committee to try to solve it," and "Committee two. It is easy to create a phrase candidate sequence such as "I have requested to make a bid decision."

FIG. 8 is a diagram showing a specific configuration of the phrase candidate table BCTBL.

The "independent word" stores a pointer at the head position in the dictionary where the independent word of the bunsetsu candidate exists.

The "adjunct word string" is composed of 2 bytes, and is an area for specifying the adjunct word string following the independent word of the bunsetsu candidate. The first 1 byte indexes the first character of the auxiliary word string on the input buffer, and the next 1 byte indexes the last character of the auxiliary word string on the input buffer. For example, in the case of the input buffer shown in FIG. 3, 26 is set in the first byte and 32 is set in the second byte to express "like". When there is no adjunct word string, it is indicated by "φ" in the figure.

The "younger brother link" links another phrase candidate starting from the same reading position as the phrase candidate.

The "child link" links a bunsetsu candidate following the bunsetsu candidate.

Note that the link means that it terminates when the value is 0.

For example, the younger brother link of the phrase candidate 0 (to the committee) is 50, and the phrase candidate 50 (to the committee) is linked. Candidate 0
The child link (to the committee) is 500 and the clause candidate is 50
Link 0 (solve). The child link of the phrase candidate 500 is 800, and the phrase candidate 800 (as illustrated) is linked. The child link of the phrase candidate 800 is 1000,
Link the phrase candidate 1000 (requested). Furthermore, it can be seen that the child link of the phrase candidate 1000 is 0, and the input reading ends there.

FIG. 9 is a diagram showing an example of sentence likelihood calculation. The sentence likelihood expresses the likelihood of a sentence that is a phrase candidate sequence,
The larger the value, the more likely it is.

The sentence likelihood is calculated by adding all the sums of the phrase likelihoods, the sums of the inter-phrase likelihoods, and the sums of the example likelihoods.

The phrase likelihood expresses the likelihood of each phrase candidate, and the larger the value, the more likely it is. As the phrase likelihood, the word likelihood described in the dictionary DIC of the independent word of the phrase is used.

The inter-phrase likelihood expresses the likelihood of connection between adjacent phrase candidates, and is fixed at a value of -20 in this embodiment. If the sentence is composed of n bunsetsu, there is (n-1) between the bunsetsus, so the sum of the inter-bunsetsu likelihoods is always -20.
(N-1).

The example likelihood is applied to the sentence likelihood for each example, which is applied between clauses. The calculation formula of the example likelihood differs between the case of the individual example and the case of the meaning example. When the semantic example is applied, the example likelihood is the value of the priority of the semantic classification used, minus five times the number of clauses skipped when applying the example. When the individual example is applied, the example likelihood is the value 20 minus five times the number of clauses skipped when applying the example.

According to FIG. 9, the sentence likelihood of “to the committee / to solve / to ask / request” is −35, and the likelihood of “to / to the committee / to solve / request / request”. Is -30, it can be seen that the "commit / request / request / request" of Example 2 having a large sentence likelihood value is converted to the first place.

The reason for this correct conversion is that the example “Apply to / consult” applied to Example 1 skips one clause “Solution”, while the example applied to Example 2 “Solution / Aim” Is
This is because the applied distance is short without jumping over other clauses.

The operation of the above embodiment will be described according to the flow.

FIG. 10 is a flow chart of a part for receiving a key input and performing a process.

Step 10-1 is a process for fetching data from the keyboard. The type of the key fetched in step 10-2 is determined, and the process branches to each key processing routine.

If it is a conversion key, the process branches to step 10-3, and in step 10-3, conversion processing of kana-kanji conversion is performed as described in detail in FIG. If the key is any other key, the process branches to step 10-4, and other processes such as insertion and deletion that are performed in a normal character processing device are performed. After that, the process branches to step 10-1.

FIG. 11 is a detailed flowchart of the "conversion process" in step 10-3.

In step 11-1, the phrase candidate creation process detailed in FIG. 12 is performed to create the phrase candidate table BCTBL.

In step 11-2, the first candidate determination process described in detail in FIG. 13 is performed.

In step 11-3, a conversion result is created and output based on the determined first candidate.

FIG. 12 is a detailed flowchart of the “bunsetsu candidate creation process” in step 11-1.

In step 12-1, the input buffer index i and the partial candidate table index j are initialized to 0.

In step 12-2, the dictionary is searched for word candidates based on the reading in the input buffer indicated by i.

In step 12-3, a morphological analysis process of analyzing an attached word string connected to the found word candidate is performed. As a result, bunsetsu candidates are obtained.

The phrase candidates obtained in step 12-4 are stored in the phrase candidate table. When storing, it is stored in the (j + 1) th entry. Also, set the necessary information. For example, a child link and a younger brother link are set for a phrase candidate in which this phrase candidate is a child or a younger brother. After storing, the value of j is incremented.

In step 12-5, a bunsetsu candidate that has not been terminated, that is, a bunsetsu candidate whose child link is not determined yet is found from the bunsetsu candidate table, and the next reading position is assigned to i.

In step 12-6, it is determined whether or not the child links of all bunsetsu candidates have been decided, and if there is any child link that has not been decided, the process branches to step 12-2. Otherwise it returns.

FIG. 13 is a detailed flowchart of the “first candidate determination process” in step 11-2.

In step 13-1, the maximum likelihood sentence likelihood is initialized to the minimum value permitted in processing, for example, -32767.

In step 13-2, one phrase candidate string is extracted from the phrase candidate table.

In step 13-3, as will be described in detail with reference to FIG. 14, an example applicable to the extracted phrase candidate sequence is searched, and the “example likelihood sum” which is the sum of example likelihoods is calculated.

In step 13-4, the sentence likelihood of the phrase candidate sequence is set to the ninth.
Calculate as shown in the figure.

In step 13-5, it is determined whether the calculated sentence likelihood is more likely than the maximum likelihood sentence likelihood, specifically, larger,
When it is larger, the maximum likelihood sentence likelihood is updated to the calculated sentence likelihood in step 13-6. In step 13-7, the current phrase candidate sequence is stored as the phrase candidate sequence corresponding to the maximum likelihood sentence likelihood.

In step 13-8, it is determined whether another phrase candidate string can be extracted from the phrase candidate table, and if it can be extracted, the process branches to step 13-2. If not, return.

As a result, the phrase candidate sequence from which the maximum likelihood sentence likelihood is calculated is determined as the first candidate.

FIG. 14 is a detailed flowchart of “calculation of example likelihood sum” in step 13-3.

In step 14-1, the example likelihood sum is first initialized to 0.

At step 14-2, one phrase is extracted as the focused phrase from the phrase candidate sequence. In the following process, the likelihood of the example related to this focused phrase will be obtained.

In step 14-3, the maximum likelihood example likelihood is initialized to 0.

In step 14-4, a phrase paired with the focused phrase is extracted from the phrase candidate string.

In step 14-5, when the pair phrase of the target phrase cannot be extracted, the processing of the further target phrase is abandoned and the process branches to step 14-8.

In step 14-6, as described in detail in FIG. 15, it is checked whether or not the example between the target phrase and the pair phrase is applied, and the example likelihood is set according to the application situation.

In step 14-7, it is checked whether or not the obtained example likelihood is more likely (ie, larger) than the maximum likelihood example likelihood, and if there is likelihood (ie, larger), the value of the example likelihood is set to the maximum likelihood example likelihood. To. Then step 14
-4 and obtain the example likelihood for another pair.

In step 14-8, since the processing of the target phrase is completed, the process moves to the next target phrase, but it is determined whether or not there is any unprocessed phrase. If not, the process returns as it is, but if it remains, the process branches to step 14-9 to add the maximum likelihood example likelihood obtained last time to the example likelihood wheel. After that, the process branches to step 14-2, and the next focused phrase is extracted.

FIG. 15 is a detailed flow chart of the “value setting of example likelihood” in step 14-6.

In step 15-1, whether or not there is an example applied to the two phrase senses of the target phrase and the pair phrase, and if there is,
Determine if it is an individual or semantic example.

When there is no application example, the process branches to step 15-2, 0 is assigned to the example likelihood, and the process returns.

If the individual example is applicable, the process branches to step 15-3, and 20 is substituted into the example likelihood in step 15-3. After that, the process branches to step 15-5.

When the meaning example is applicable, the process branches to step 15-4, and the priority of the meaning classification used in the meaning example is substituted as the example likelihood. After that, the process branches to step 15-5.

In step 15-5, the distance between two clauses, that is,
The number of bunsetsus to jump over is calculated, and the value is multiplied by 5 and added to the example likelihood. If there are two adjacent clauses, the jumping clause is 0, so 0 is added to the example likelihood.

In step 15-6, it is determined whether or not the calculated value of the example likelihood is negative, and if negative, the value is corrected to 0 in step 15-7, and the process returns. When it is not negative, it returns without changing the value.

[Other Embodiments] In the above description, the example dictionary is described as a pair of co-occurring two words (or classifications), but the same processing is performed even if it is a set of three or more generally n sets. can do.

Also, the likelihood calculation is merely an example, and if the distance between multiple clauses to which the example is applied is considered in the process of likelihood calculation, even if another calculation method is used, It does not impair the gist of the present invention.

[Effects of the Invention] As is clear from the above description, according to the present invention, when a plurality of examples are applicable to the reading sequence input by the operator, the distance of two phrases to which the example is applied (the number of skipped phrases). According to this, it is determined which example should be applied, so that more appropriate conversion can be performed. This makes it possible to realize a comfortable character processing device with a high conversion rate.

[Brief description of drawings]

FIG. 1 is a block diagram of the overall configuration of the present invention, FIG. 2 is a diagram showing a conversion example of kana-kanji conversion in the present invention, and FIG. 3 is a configuration of an input buffer IBUF and an output buffer OBUF in the present invention. FIG. 4, FIG. 4 is a diagram showing a structure of a kana-kanji conversion dictionary DIC in the present invention, FIG. 5 is a diagram showing a structure of an example dictionary YDIC in the present invention, and FIG. 6 is a meaning classification priority in the present invention. The figure which showed the structure of the meaning classification priority table PRTBL which described the relationship, FIG. 7 is the figure which showed notionally the memory content of the clause candidate table BCTBL in this invention, and FIG. 8 is the clause candidate table BCTBL in this invention. FIG. 9 is a diagram showing the configuration of FIG. 9, FIG. 9 is a diagram showing a method of likelihood calculation in the present invention, and FIGS. DISK ... External storage unit CPU ... Microprocessor ROM ... Read-only memory RAM ... Random access memory IBUF ... Input buffer OBUF ... Output buffer DIC ... Kana-Kanji conversion dictionary YDIC ... Example dictionary PRTBL ... Meaning classification priority table BCTBL ... Clause candidate table

Claims (1)

[Claims] 1. An input means for inputting a kana character string, a word dictionary means for storing a reading of a word, a notation and a semantic classification in association with each other, a word notation, and a word used in combination with the word. Example dictionary means for storing as an example in association with the notation or the semantic classification, and a candidate creation for creating a notation candidate for reading the kana character string input by the input means by referring to the word dictionary means Means, and for the notation candidates for the same reading created by the candidate creating means, the example stored in the example dictionary means gives priority to the corresponding candidate together with the word used together, and when there are a plurality of the applicable candidates, A display priority determining means for determining a display priority of a notation candidate by giving priority to a candidate corresponding to an example in which a distance between the candidate and the word used in combination is short in a sentence, A character processing device, comprising: candidate display means for displaying a notation candidate created by the candidate creating means based on the display priority determined by the display priority determining means.