JPH03167657A - Character processor - Google Patents

Abstract

PURPOSE:To improve the conversion efficiency of examples by adopting the semantics examples of higher priorities which are described in a word dictionary to decide a 1st candidate for KANA (Japanese syllabary)/KANJI (Chinese character) conversion when the conflict of semantics examples occurs for an input read string. CONSTITUTION:A KANA character string, e.g., 'YAOYA NI ITSUTA', etc., inputted via a keyboard KB is stored in a key input buffer area IBUF of a RAM under the control of a CPU. Then the KANA characters like 'YAOYA', etc., are converted into KANJI by reference to a KANA/KANJI conversion dictionary area DIC. The area DIC stores the semantic sorts like the place, the human, etc., of the converted KANJI and the priorities corresponding to the semantic sorts as well as the converted KANJI itself as the stepped likelihood value. An examples dictionary area YDIC is referred to in response to the likelihood value so that an example like 'YAOYA NI ITTA (gone)' not 'YAOYA NI ITTA (said)' is decided as a 1st conversion candidate if 'YAOYA' has the highest priority as a place. Thus the proper KANA/KANJI conversion is carried out for the homonym KANA character strings with high conversion efficiency.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a character processing device for inputting a sentence containing kanji and kana through kana-kanji conversion.

[Prior art] Currently, character processing devices such as Japanese word processors are...
It is common to use kana-kanji conversion to input sentences containing sweet characters and kana.

Kana-kanji conversion converts the input pronunciation sequence into kanji by referring to a dictionary. In dictionaries, part-of-speech information such as noun, sa-hen noun, adverb, adjective, and adjective verb is written for each word, and kana-kanji conversion analyzes the pronunciation sequence to create possible clause candidates and combines them. conversion candidates are determined and presented in order of likelihood. The operator selects a desired candidate from among the presented conversion candidates.

For example, for the pronunciation sequence ``Shinbunya ni Ita'', ``Newspaper shop'', ``Newspaper store'', ``Newspaper shop'', ``Newspaper'', ``Yani'' are used.
``to the arrow''``arrow''``similar''``said'' ``went j
Clause candidates such as ``I told the newspaper shop'' and ``I went to the newspaper shop'', which are a combination of these, are output and displayed as conversion candidates. Since it cannot be said which of "said" and "done" has a higher frequency, it is not necessarily converted as the first candidate.

Therefore, a method called example conversion has been proposed to increase the conversion rate. Example conversion is a method in which bears (examples) of co-occurring words are registered in advance in an example dictionary, the example dictionary is referred to at the time of conversion, and the first candidate is changed according to the example in the example dictionary. For example, if you store the pattern ``to/consult with a committee'' in the example dictionary, then ``consult with the committee'' will be converted as the first option when inputting ``Iinkai ni Hatta''. Make it.

The above example is an example of an individual usage example that describes the relationship between words, but semantic usage examples that describe the relationship with semantic classification have also been proposed. For example, in the context of "I went to...", when "..." represents a place, "I went to..."
The notation ``said'' is plausible, and when ``r~'' represents a human being, the notation ``said'' is plausible. In order to perform this conversion correctly, examples of meaning such as ``to say to a person'' and ``to go to a place'' are registered in advance in the example dictionary. Note that words such as ``human'' and ``place'' are semantic classifications, and all words in the dictionary are described as being ``human'', ``place'', or not. With this mechanism, it is possible to correctly convert sentences such as "I told the mayor" and "I went to the city hall."However, conventional example conversion techniques are difficult to use when multiple examples are applicable, that is, when there are conflicts, which example is used? There was no specific specification as to which should be prioritized. There is only a guideline to give priority to the individual example when there is a slight conflict between the individual example and the semantic example, but in other situations it was unclear which example should be given priority.

Therefore, since the first candidate was determined appropriately by adopting an appropriate usage example, the conversion rate in delicate situations was excellent.The present invention describes the priority for the semantic classification stored in the word dictionary. , when multiple semantic examples compete during conversion, the semantic example using a high-priority semantic classification is preferentially adopted and the first candidate is determined, aiming to improve the conversion rate of example conversion. be.

[Example] 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 that performs calculations, logical judgments, etc. for character processing, and is connected to an address bus AB, a control bus CB, and a data path D.
B to control each component connected to those buses.

Address bus AB transfers address signals indicating the components to be controlled by the microprocessor CPU. The control bus CB transfers and applies control signals for each component to be controlled by the microprocessor CPU. The data path DB transfers data between each component device.

Next, ROM is a read-only fixed memory, and the first
Microprocessor CP, which will be described later in FIGS. 2 to 19.
The control procedure by U is memorized.

Further, the RAM is a writable random access memory having a configuration of 1 word and 16 bits, and is used for temporary storage of various data from each component. I BUF is an input buffer that stores key data entered manually, and O
BUF is an output buffer that temporarily stores the results of kana-kanji conversion. DIC is a dictionary for performing kana-kanji conversion. YDIC is an example dictionary that describes the co-occurrence relationship between certain words and unclassified words. BCTBL is a phrase candidate table that stores phrase candidates that are currently being converted. TBU
F is a text buffer in which text data being input and edited is stored. DOBUF is a homophone buffer, and when a homophone candidate exists for a character stored in the text buffer TBUF, that candidate is stored.

KB is a keyboard, which includes character symbol input keys such as alphabetical hirakana keys and katakana keys, and various functions for instructing various functions to this character processing device, such as conversion keys, next candidate keys, and selection keys. Comes with a key.

The DISK is an external storage section for storing document data, and stores documents created on the text buffer TBUF, and the stored documents can be called up when necessary by instructions from the keyboard.

CR is a cursor register. The CPU can read and write the contents of the cursor register.

A CRT controller CRTC, which will be 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 patterns to be displayed. When displaying text contents, a display pattern is created in DBUF according to the contents of text buffer TBUF, and the text is displayed.

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

Further, a CRT is a display device using a cathode ray tube or the like, and a CRT controller controls the dot-configured display pattern and cursor display on the display device CRT.

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

The character processing device of the present invention, which is composed of each of these components, operates in response to various human inputs from the keyboard KB, and when an input from the keyboard KB is supplied, an interact signal is first sent to the microprocessor CP.
The microprocessor CPU reads various control signals stored in the ROM, and various controls are performed in accordance with these control signals.

FIG. 2 is a diagram showing an example of conversion by the apparatus of the present invention. In the figure, r/J means that the conversion key is pressed there.

When "I went to Yaoya" is input, the first option is converted to "I went to a greengrocer." This is written in the example dictionary as "
An example of the meaning "to/go to a place" has been registered,
This is because "place" is registered as a semantic classification for "greengrocer." Without this kind of example conversion mechanism, ``I went'' would be replaced with ``I went,'' ``I said,'' and ``I needed.''
Because there are homophones such as , there is no guarantee that it will be converted correctly. Furthermore, the second option is converted to "I told the greengrocer." This is written in the usage dictionary as ``to ``human''/
This is because the semantic example of "say" is registered, and "human" is registered as the semantic classification of "greengrocer".

Regarding "greengrocer", the semantic classification is "human" rather than "human".
It is converted like this because the priority is given to "Location".

The next example is when you enter "I went to Shinbunya." The word ``newspaper shop'' also has the meaning ``human'' in it.
Both "Location" are registered, but in the case of "Newspaper Store", "Human" is given higher priority than "Location", so unlike the first example, the first candidate is "Newspaper". The second option is ``I went to the newspaper store.''

FIG. 3 is a diagram showing an example of learning of semantic classification by the apparatus of the present invention. 3-1 shows the initial screen.

3-2 shows the screen when the reading string ``Shinbunya niita'' is input. The cursor is displayed next to the input reading sequence. If you press the conversion key here, the screen 3-3 will appear. In 3-3, the reading sequence ``I went to the Shinbunya'' is converted to ``I told the newspaperman.'' ``Newspaper store J has ``human'' and ``place'' as semantic classifications, and ``human''
There are examples such as ``to/say'' and ``to/go to <place>,'' but for ``newspaper shop,'' the usage example for ``human'', which has the highest priority, is applied and converted. Assuming that the operator desires the conversion ``I went to the newspaper store'', when he presses the next candidate key, the screen 3-4 appears. Here, conversion candidates for “ita” are displayed. The first candidate is "said" and the second candidate is "did." The current candidate is the second candidate “
"I went," and 2 is highlighted. If you press the selection key here, the screen 3-5 will appear.

The second candidate "I went" is determined and stored in the text data, and at the same time, the semantic classification "location" of "newspaper shop" is learned and the priority is improved. If you enter "I went to Shinbunya" again, the screen shown in 3-6 will appear. If you press the conversion key again, the screen shown in 3-7 will appear, and this time the semantic classification ``location'' has been learned, so it will be applied preferentially, and ``I went to the newspaper store'' will be displayed as the first option. .

Figure 4 shows large buffer IBUF and output buffer OBUF.
FIG.

Both IBUF and OBUF have the same configuration. The first two bytes are size information for each buffer, and contain a value obtained by subtracting 1 from the number of characters stored in the buffer and multiplying by 2. r/J at the end of the large buffer means that the conversion key was pressed there. Each character consists of 2 bytes and is stored in JIS X 0208 code or the like.

FIG. 5 is a diagram showing the configuration of the dictionary DIC.

It consists of fields for ``pronunciation,''``orthography,'' ``part of speech,'' ``word likelihood,'' and ``semantic classification.''

"Yomi" means the reading of the word, "notation" means the notation of the word, "
``Part of speech'' stores the part of speech of a word. "Word likelihood" is information indicating the likelihood of the word itself, such as frequency information, from 1 to
Stored as a value of 5.

A likelihood value of 5 means the most likely, and the smaller the value, the more doubtful it is. Since the likelihood value O means that it is completely unthinkable, it does not exist as a word likelihood value.

``Semantic classification'' shows the meaning classification of the word, ``Organization'' ``
It is generally described in multiple terms, such as place》《human being》.

Furthermore, for each semantic classification, the likelihood of interpreting the word as belonging to that semantic classification (priority information) is determined.
is stored as a value of 1 to 5. Priority 5 means the most likely, and the smaller the value, the more doubtful it is. Priority O means that it cannot be considered at all, so it does not exist as a value. However, semantic classification is only described when the word is a noun.

FIG. 6 is a diagram showing the structure of the example dictionary @YDIC. "
It consists of fields for 1st word, 2nd word, particle, and priority.

"First word" and "Second word" describe pairs of words that co-occur. If it is not an individual example but a semantic example, the semantic classification is written instead of the word.

"Particle" describes the particle that connects the bare words.

"Priority" indicates the likelihood of the example from 1 to 1.
Stored as a value of 5. A likelihood value of 5 means the most likely, and the smaller the value, the more doubtful it is. Since the likelihood value O means that it is completely unthinkable, it does not exist as a priority value.

In other words, the example shown in the diagram is "at <<location>>/line<"
(Priority 2) "Tell a human being" (Priority 2) "Consult with a committee" (Priority 5) "Raise an animal" (Priority 2)

Note that semantic examples are generally given lower priority than individual examples. By doing this, the principle that ``individual usage takes precedence over semantic usage'' will be naturally realized.

FIG. 7 is a diagram showing the concept of the phrase candidate table BCTBL. The phrase candidate table is a pinary tree representation of possible phrase candidates as a result of analyzing the input reading. In the figure, horizontal lines mean child pointers, and vertical lines mean younger brother pointers. The younger brother pointer links other clause candidates (usually shorter candidates) starting from a certain reading position, and the child pointer links the clause candidates that follow that clause.

The first part of the input pronunciation string ``Shinbunya ni Ita'' has interpretations such as ``newspaper shop'', ``newspaper shop'', ``newspaper shop'', and ``newspaper'', and these are linked in order by younger brother pointers (warp springs). There is.

The phrase that follows ``to the newspaper shop'' is thought to be ``said.''
Linked by child pointer.

Once a phrase candidate table like this is created, it can be summarized as ``I told the newspaper shop,'' ``I went to the newspaper shop,'' or ``I said something similar to the newspaper.''
``It is easy to create a phrase candidate string such as ``newspaper yani go datta j''.

FIG. 8 is a diagram showing the specific structure of the clause candidate table BCTEL.

"Independent word" stores a pointer to the top position in the dictionary where the independent word of the bunsetsu candidate exists.

The "adjunct word string" consists of 2 bytes, and is an area for specifying the adjunct word string that follows the independent word of the bunsetsu candidate. The first byte indexes the first character of the adjunct word string on the input buffer, and the next one byte indexes the last character of the adjunct word string on the input buffer. For example, when using the large buffer shown in Figure 4, to express ``ta'', the first byte contains 14
, set 16 in the second byte. When an adjunct word string does not exist, it is indicated by "φ" in the figure.

"Little brother link" links another clause candidate that starts from the same reading position as the clause candidate.

A "child link" links a clause candidate that follows the clause candidate.

Note that the link is terminated when the value is O.

For example, the younger brother link of clause candidate O (to a newspaper shop) is 50, which links clause candidate 50 (newspaper shop). Clause candidate O
The child link of (to the newspaper shop) is 500, and there are 50 phrase candidates.
Link 0 (said). The child link of the bunsetsu candidate 500 is O, and it can be seen that the input reading ends there.

FIG. 9 is a diagram showing the structure of the text buffer TBUF.

A text buffer consists of a sequence of characters, each character consisting of 2 bytes. The MSB of each letter is a homophone flag, O means a normal letter, and l means a homophone. The remaining 15 bits represent the character code in the case of a normal character, and the homophone number in the case of a homophone. For example, the JIS X 0208 code is used as the character code. The homophone number is a number indicating which homophone is on the homophone buffer DOBUF shown in FIG.

FIG. 10 is a diagram showing a smaller configuration of the homophone buffer DOBUF. Each homophone is identified by a homophone number.

Each homophone is ``pronunciation j ``Total number of candidates'' ``Current candidate number''
Consists of "I-th Candidate Information".

"Yomi" stores the reading of its homophone.

"Total number of candidates" stores the total number of candidates included in the same sound.

The "current candidate number" stores the currently displayed candidate number of the homophone. Since the first candidate is displayed as the initial value immediately after conversion, "1" is stored.

The "i-th candidate information" stores the "notation", "word address", and "applied meaning classification" of each candidate.

"Notation" stores the notation of the candidate.

``Word Address'' and ``Applicable Semantic Classification'' store the existence address of a word that matches the semantic classification described by the semantic example and its semantic classification code when there is a semantic example applied to the candidate. Usually, there are multiple usage examples applicable to the candidate, so multiple classifications are included in the ``existence address'' and ``applicable meaning classification.''

For example, if it is converted to "I told the newspaperman," the candidate "
For "Said," the address of "newspaper shop" is stored as the "word address," and "human" is stored as the "applied meaning classification." If it is converted to ``I went to the newspaper room,'' the candidate ``I went to the newspaper room'' will be changed to ``Word Address.''
``Location'' is stored as ``Newspaper Store'' and ``Applicable Semantic Classification.''

FIG. 11 is a diagram showing an example of calculating sentence likelihood.

Sentence likelihood expresses the likelihood of a sentence that is a clause candidate sequence, and the larger the value, the more likely it is.

The sentence likelihood is calculated by adding the sum of clause likelihoods, the sum of inter-clause likelihoods, and the sum of example likelihoods.

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

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

The example likelihood is added to the sentence likelihood for each example applied between each clause. The calculation formula for example likelihood differs depending on whether it is an individual example or a semantic example. In the case of individual examples, the example likelihood is obtained by subtracting twice the number of clauses to be skipped when applying the example from four times the priority described in each example. In the case of semantic examples, the priority written in each example is multiplied by 4 times the priority of the semantic classification, divided by 5, minus twice the number of clauses to be skipped when applying the example. is the example likelihood.

According to Figure 11, the sentence likelihood of "I went to the newspaper store" is -
8.2, and the sentence likelihood of "I told the newspaperman" is -6.
6, it can be seen that "I told the newspaperman" in Example 2, which has a large value of sentence likelihood, is converted to the first place.

The reason why it is converted correctly in this way is that in example 1, "newspaper shop"
In Example 2, the priority of the semantic category "Place" that is interpreted is 3, whereas in Example 2, the semantic category "Human" is interpreted as "Newspaper store".
This is because the priority is 4, and a sentence that interprets "newspaperman" as "human" is more advantageously converted.

Note that if the priorities of "person" and "place" are reversed, the first and second candidates will be converted in the opposite order.

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

FIG. 12 is a flowchart of the part that takes in key human power and performs processing.

Step 12-1 is a process of capturing data from the keyboard. In step 12-2, the type of the retrieved key is determined, and the process branches to a processing routine for each key.

If it is a conversion key, the process branches to step 12-3, and in step 12-3, conversion processing for kana-kanji conversion is performed as detailed in FIG. 13. If it is the next candidate key, the next candidate process detailed in FIG. 18 is performed in step 12-4. If it is a selection key, step l
At step 2-5, the selection process detailed in FIG. 19 is performed.

If it is any other key, the process branches to step 12-6, and other processing such as insertion, deletion, etc. performed in a normal character processing device is performed. After that, it branches to step 12-1.

FIG. 13 is a detailed flowchart of the "conversion process" in step 12-3.

In step 13-1, a phrase candidate creation process detailed in FIG. 14 is performed to create a phrase candidate table BCTBL.

In step 13-2, a first candidate determination process detailed in FIG. 15 is performed.

In step 13-3, a conversion result is created and output based on the determined ↓-th candidate.

FIG. 14 is a detailed flowchart of the "phrase candidate creation process" in step l3-1.

In step 14-l, the input buffer index i and the phrase candidate table index j are initialized to O.

In step 14-2, a dictionary is searched for word candidates based on the pronunciation in the human cover text indicated by i.

In step 14-3, morphological analysis processing is performed to analyze adjunct word strings connected to the found word candidates. As a result, phrase candidates are obtained.

The clause candidates obtained in step 14-4 are stored in the clause candidate table. When storing, it is stored in the j-thousand-1st entry. Also, set the necessary information. For example, for a clause candidate that has this clause candidate as a child or younger brother, a child link and a younger brother nonku are set. Store? Count up the value of &j.

In step 14-5, find an unterminated phrase candidate from the phrase candidate table, that is, a phrase candidate whose child link has not yet been determined, and determine the next reading position i.
Assign to .

In step 14-6, it is determined whether the child links of all clause candidates have been determined, and if there are any child links that have not been determined, the process branches to step 14-2. Otherwise, return.

FIG. 15 is a detailed flowchart of the "first candidate determination process" in step l3-2.

In step 15-1, the maximum sentence likelihood is initialized to the minimum value allowed for processing, for example, -32767.

In step 15-2, one phrase candidate string is extracted from the phrase candidate table.

In step 15-3, as detailed in FIG. 16, an example applicable to the extracted sentence candidate string is searched, and a "sum of example likelihoods" which is a sum of example likelihoods is calculated.

In step 15-4, the sentence likelihood of the phrase candidate string is
Calculate as shown in Figure 1.

In step l5-5, it is determined whether the calculated sentence likelihood is more likely than the maximum likelihood sentence likelihood, specifically, whether it is larger;
If it is larger, the maximum likelihood sentence likelihood is updated to the calculated sentence likelihood in step 15-6. Further, in step l5-7, the current phrase candidate string is stored as a phrase candidate string corresponding to the maximum sentence likelihood.

In step 15-8, 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 15-2. Otherwise, return.

As a result, the phrase candidate string that is the basis for calculating the maximum likelihood sentence likelihood is determined as the first candidate.

FIG. 16 is a detailed flowchart of step 15-3, ``Calculation of example likelihood sum''.

In step 16-1, the example likelihood sum is first initialized to O.

In step 16-2, one clause is extracted from the clause candidate string as a clause of interest. In the following processing, the likelihood of examples related to this phrase of interest will be calculated.

In step 16-3, the maximum likelihood example likelihood is initialized to O.

In step 16-4, phrases that are bare of the phrase of interest are extracted from the phrase candidate string.

In step 16-5, if the bare clause of the target clause cannot be extracted, processing of the target clause is given up and the process branches to step 16-8.

In step 16-6, as detailed in FIG. 17, it is checked whether the example between the target phrase and the paired phrase is applied, and the example likelihood is set depending on the application status.

In step 16-7, it is checked whether the obtained example likelihood is more likely (i.e., larger) than the maximum likelihood example likelihood, and if it is likely (i.e., larger), the value of the example likelihood Assign to . Then step l6
-4, and find the example likelihood for another bear.

In step 16-8, since the processing of the target phrase is completed, processing moves on to the next target phrase, and it is determined whether there are any unprocessed phrases remaining. If there are no remaining examples, the process returns as is, but if there are any, the process branches to step 16-9, and the previously determined maximum likelihood example likelihood is added to the sum of example likelihoods. Thereafter, the process branches to step 16-2, and the next phrase of interest is extracted.

FIG. 17 is a detailed flowchart of step 16-6, ``setting example likelihood value.''

In step 17-1, it is determined whether there is an example that applies between the two phrases of interest and a pair of phrases, and if so, whether it is an individual example or a semantic example.

If there is no applied example, the process branches to step 17-2, O is substituted for the example likelihood, and the process returns.

When the individual example is applied, branch to step ■7-3,
The priority of the example multiplied by 4 is assigned as the example likelihood. After that, the process branches to step 17-5.

When a semantic example is applied, the process branches to step 17-4, and the value obtained by multiplying the priority of the example by 4, further multiplied by the priority of the semantic classification, and dividing by 5 is substituted as the example likelihood.

Thereafter, the process branches to step 17-5.

In step 17-5, the distance between two clauses, that is,
Find the number of phrases to skip, multiply that value by 2, and subtract it from the example likelihood. If the two clauses are adjacent, the clause to be skipped is O, so O is subtracted from the example likelihood.

In step 17-6, it is determined whether or not the calculated example likelihood value is negative, and if it is negative, it is corrected to be O in step 17-7, and the process returns. If it is not negative, return without changing the value.

FIG. 18 is a detailed flowchart of the "next candidate processing" in step 12-4.

In step 18-1, the homophone number of the homophone to be viewed as the next candidate is determined from the text buffer TBUF.

In step 18-2, the position of the homophone buffer is determined from the homophone number, and the current candidate number is counted up.

In step 18-3, a list of candidates is displayed.

FIG. 19 is a detailed flowchart of the "selection process" in step 12-5.

In step 19-1, the homophone number of the homophone to be selected is obtained from the text buffer TBUF.

In step 19-2, the position of the homophone buffer is determined from the homophone number, the notation is extracted from the candidate information indicated by the current candidate number, and is set as a confirmed character in the text buffer TBUF.

In step 19-3, the word address and applied semantic classification are similarly determined from the candidate information indicated by the current candidate number, and the priority of the semantic classification of the indicated word is counted up.

In step 19-4, the word address and applied semantic classification are similarly determined from the candidate information indicated by the first candidate, and the priority of the semantic classification of the indicated word is counted down.

[Other Examples In the above explanation, the example dictionary has been explained as a pair of two co-occurring words (or categories), but it can be processed in the same way even if it is a set of three or more generally a set of n words. can do.

Furthermore, the calculation of the likelihood is merely an example, and if the priority given to the example is taken into consideration in the process of likelihood calculation, even if another calculation method is used, the present invention is particularly effective. This does not detract from the purpose.

Furthermore, the semantic classification learning mechanism is configured to directly manipulate the priority of the semantic classification written in each word. However, in addition to this method, it is possible to have learning bits corresponding to each semantic classification of words and to manipulate the learning bits without detracting from the spirit of the present invention.

[Effects of the Invention] As is clear from the above description, according to the present invention, when a plurality of semantic usage examples can be applied to the pronunciation string manually created by the operator, the semantic classification described for each word in the word dictionary is Since it is determined which semantic example to apply based on the priority information, more appropriate conversion can be performed.

Furthermore, since the priority information for semantic classification can be changed through learning, even operators who use words with special meanings can improve the conversion to their own preference.

This makes it possible to realize a comfortable character processing device with a high conversion rate.

[Brief explanation of the drawing]

Figure 1 is a block diagram of the overall configuration of the present invention, Figure 2 is a diagram showing an example of kana-keiji conversion in the present invention, and Figure 3 is a diagram showing the learning effect of priority of usage in the present invention. , FIG. 4 is a diagram showing the configuration of the input buffer IBUF and output buffer OBUF in the present invention, FIG. 5 is a diagram showing the configuration of the kana-kanji conversion dictionary DIC in the present invention, and FIG. 6 is an example of use in the present invention. FIG. 7 is a diagram conceptually showing the storage contents of the clause candidate table BCTBL in the present invention; FIG. 8 is a diagram showing the structure of the clause candidate table BCTBL in the present invention; Figure 9 is a diagram showing the configuration of the text buffer TBUF in the present invention, Figure 10 is a diagram showing the configuration of the homophone buffer DOBUF in the present invention, and Figure 11 is a diagram showing the likelihood calculation method in the present invention. , FIGS. 12 to 19 are flowcharts showing the operation of the character processing apparatus of the present invention. DISK...External storage CPU...Microprocessor ROM
...Read-only memory RAM ...Random access memory I BUF ...Large buffer OBUF ...Output buffer DIC ...Kana-kanji conversion dictionary YDIC
...Example dictionary BCTBL ...Bunsetsu candidate table TBUF
・・・Text buffer DOBUF ・・・Homophone buffer I went to Shinbunya〃→ I went to the newspaper shop 4 Illustrated Kabanofa's horizontal output Panofa's composition 2 bytes 1 2 bytes 0 Part 5 Figure 6 Figure input ``I went to Shinbunya'' Independent words that went 8 Figure attached 9 Meditation 2 brother link Child link 0 Ordinary character 1 Homophone 10 Figure Homophone number O l 2 Funeral 11 Figure (sentence likelihood) 1 Σ (bunsetsu luminosity) ↓Σ (t degree between clauses) book Σ (
Example likelihood) (clause likelihood) = R: i-word likelihood (inter-clause likelihood) - 20 (fll1) 1Calculation of the sentence likelihood of ``to/went to Shinsekiya'' Clause likelihood of ``to new Illm'' 1. Word likelihood of “New 1! l-ya” - 2. Clause likelihood of “Go ko” 1. Word likelihood of “Go j” -
Sum of likelihoods between 5 clauses - 1 (2 - 1) x 20 = -20 "New!
Ill shop / line <. Example of usage of ``Dominance -2x4x3;510=4 Therefore, sentence likelihood -2 + 5 - 20 + 4.8yo82 (Example 2) Calculating the sentence likelihood of ``I went to the newspaper store'' Clause ``To the newspaper store'' Likelihood 1 “Newspaper shop.; word likelihood -2
``Bunsetsu likelihood of ``3'' 1 Word likelihood of ``say'' = 5 Sum of inter-clause likelihoods - (2-1) x 20 - 20 ``New III shop/g UJ's''A example 1jy =2x4x4+5-0
1t for -6 = 2 + 5 - 20 = 6. 4=-6. 6
Figure 12 Funeral 13 Figure 14 Figure 17 Figure ○ Death B

Claims (1)

[Claims] 1. An input means for inputting a pronunciation sequence; a dictionary storing pronunciations, notations, and semantic categories in correspondence; and an example dictionary describing co-occurrence relationships between a plurality of words or semantic categories; a converting means for converting the input pronunciation sequence into notation based on the dictionary and the example dictionary, priority information is described for each word in the semantic classification of the dictionary, and the converting means converts the semantic example into a notation. A character processing device characterized in that, when applied, a first candidate is determined by preferentially applying an example using a meaning classification with a high priority. 2. An input means for inputting a pronunciation sequence; a dictionary that stores pronunciations, notations, and semantic categories in correspondence; an example dictionary that describes co-occurrence relationships between a plurality of words or semantic categories; a conversion means for converting an input pronunciation sequence into a notation based on the notation; a next candidate means for displaying the next candidate when the first candidate of the conversion result is not the notation desired; and a next candidate means that displays the next candidate if the displayed candidate is the notation desired by the user. When applying the semantic examples, the conversion means selects and finalizes the candidates, and priority information is described for each word in the meaning classification of the dictionary, and when applying the semantic example, the conversion means selects and confirms the candidates. A character characterized in that the first candidate is determined by preferentially applying examples using semantic classification, and when the selection means selects a candidate below the next candidate, the priority of the semantic classification of the word is changed. Processing equipment.