JPH0283775A - Character input device and method - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION Field of Industrial Application The present invention relates to a character input device and method for converting a Japanese sentence specified by reading into a Japanese character string including kanji and kana and inputting it in a word processor, calculator, etc. It is something.

BACKGROUND OF THE INVENTION Many conventional character input devices and methods have been devised as kana-kanji conversion devices and methods, and in recent years, devices that utilize "related words" have been announced in particular.

"Related words" are word pairs that co-occur in a sentence.Using the fact that "warm" and "cooking" are related words, when "cooking" is present in the context, "warm"'' can be converted and output with priority over ``warm'' in kana-kanji conversion.

An example of such a conventional character input device and method is "Solid kana-kanji conversion using collocation analysis" (Information Processing Society of Japan Japanese Document Processing Study Group 21-2 1985
5/8) etc. have been announced. Such conventional character input devices and methods will be briefly described below.

FIG. 5 is a block diagram showing the configuration of a conventional kana-kanji conversion device.

It is a keyboard device equipped with kana keys of 50 syllables for inputting Japanese pronunciations as codes.

Reference numeral 2 denotes an input character string temporary storage section, which sequentially records character codes input through the kana input section 1 and temporarily stores them as a reading code string.

3 is a kana-kanji conversion unit, which converts the pronunciation in the input character string temporary storage unit 2 into a corresponding conversion candidate of kanji and kana based on a comparison with an internally held word dictionary and outputs it. At this time, the output conversion candidates include the matching length of the word reading,
Also, the output priority number is incremented and output based on the order in which words are stored in the dictionary. - The longer the matching length, the smaller the number, and if they are equal, the earlier the storage order in the dictionary, the smaller the number.

4 is a conversion instruction input unit, which is realized by a keyboard device equipped with conversion keys. The operator inputs the conversion instruction input section 4.
When the conversion key is pressed, a "conversion instruction signal" for instructing conversion start is output to the kana-kanji conversion unit 3.

Reference numeral 5 denotes a conversion candidate temporary storage section which stores conversion candidate word strings of Kanji-kana mixed strings outputted by the kana-kanji conversion section 3 in the order of output priority numbers.

Reference numeral 6 denotes a conversion candidate display unit that displays the conversion candidate word strings in the conversion candidate temporary storage unit S in association with display numbers on the Japanese language display device.

Reference numeral 7 denotes a conversion candidate selection section, which outputs the conversion candidate word string stored in the conversion candidate temporary storage section 5 to the conversion candidate display section 6 in correspondence with the display number. The conversion candidate selection unit 7 accepts the “next candidate output signal” and the “selection number instruction signal” as input, and in the case of the former, the conversion candidate display unit 6 displays the next conversion candidate word string from the conversion candidate temporary storage unit 5. In the latter case, the conversion candidate word string associated with the corresponding display number displayed on the conversion candidate display section 6 is determined.

Reference numeral 8 denotes a selection instruction input section, which is a keyboard device having a "next candidate key" and a "number key" for selecting candidates. The selection instruction input section 8 outputs a "next candidate output signal" and a "selection number instruction signal" to the conversion candidate selection section 7 in response to these key presses, respectively.

9 is a related word dictionary, which stores word pairs that often co-occur in sentences in association with each other. The memory item is the notation of both words, and the conjugated word has only the stem part. Furthermore, the records in the related word dictionary 9 are sorted by the notation of one of the words (headword).

Reference numeral 10 denotes a related word search section, which searches the related word dictionary 9 for records having the conversion candidate words in the conversion candidate temporary storage section 5 as headings, and extracts related conversion candidate words.

Reference numeral 11 denotes a related word priority order change unit that changes the output priority order of conversion candidate words in the conversion candidate temporary storage unit 5 in accordance with the processing result of the related word search unit 10. That is, the output priority of words that are related to each other among the conversion candidate words is increased.

Reference numeral 12 denotes a text storage unit that stores the conversion candidate word string selected by the conversion candidate selection unit 11 as a confirmed text.

Reference numeral 13 denotes a text display unit, which is a display device that displays text character notation strings in the text storage unit 12.

Reference numeral 14 denotes an input position changing unit, which is equipped with four-direction cursor movement keys, and changes the character input position in the text in the text storage unit 12 and the conversion candidate display position in the conversion candidate display unit 6 in response to a key press instruction.

FIG. 6 is a flowchart illustrating the procedure of character input processing in a conventional character input method. Below, the operations are [Reading input], [Change input position], [Create conversion candidate],
This will be explained separately in [Conversion candidate selection].

[Reader Kako 601: Reading input stage The operator inputs the reading of the character intended to be input into the kana input section 1.
Instructions are given in order by pressing the corresponding Japanese syllabary keys above. 1
Each time a key is pressed, the reading code corresponding to each key is added and stored in the input character string temporary storage section 2 in order. As soon as the characters are stored, the system waits for input again, and when the reading is input again, the reading input this time is concatenated with the reading input previously, and the continuation of the input reading string is stored in the input character string temporary storage section 2. is formed.

[Input position change] 602: Input position change stage When the operator presses the cursor key on the input position change unit 14 with the intention of changing the input position in the sentence, the conversion candidate selection unit 7 changes the input position change unit 14. According to instructions from
Confirms the currently displayed candidates. After that, the character input position in the text storage section 12''2 is changed, and the change is displayed on the conversion candidate display section 16.

[Conversion candidate creation] 603: Conversion instruction step When the operator presses the "conversion key" of the conversion instruction input section 4 to instruct the start of the process, the conversion instruction input section 4
outputs the corresponding "conversion instruction signal" to the kana-kanji converter 3. The kana-kanji conversion unit 3 starts conversion processing for the input reading in the input character string temporary storage unit 2 at that time.

604: Conversion candidate generation stage The kana-kanji conversion unit 3 searches for a word that matches the beginning of the kana reading string in the input character string temporary storage unit 2 in the word dictionary held within the kana-kanji conversion unit 3 and can interpret it. The conversion candidate word string is stored in the conversion candidate temporary storage section 5. Interpretable means that the input pronunciation is applied to the word and it is composed of an independent word and an attached word.
This refers to cases in which a phrase matches as a repetition of a phrase. At this time, if multiple clause breaks are possible, interpret the first two clauses at a time and give priority to the one that can be interpreted with the longest reading.

The priority order among homophones is the same as the order in which they are stored in the word dictionary.

605: Related word search stage When the conversion candidate word string is stored in the conversion candidate temporary storage section 5, the related word search section 10 searches the related word dictionary 9 for the words stored in the conversion candidate temporary storage section 5.

If the corresponding word does not exist, control is transferred to 607. If it exists, it is further checked whether the 91-word of the searched related word dictionary record pair also exists in the conversion candidate temporary storage section 5. If it exists, the process proceeds to the output priority change step 606, and if not, the control moves to the conversion candidate display step 607.

606: Output priority order change stage If a related word pair is stored in the conversion candidate temporary storage section 5, the output order of the corresponding word in the conversion candidate temporary storage section 5 among homophones is set to 1. When there are multiple related word pairs, the shorter the distance between the words forming the related word pair, the higher the priority is given, and if the words are the same, processing is applied in the order of input.

+1 607: Conversion candidate display stage The conversion candidate selection unit 7 creates conversion candidates by combining words with lower output priority among the conversion candidate words stored in the conversion candidate temporary storage unit 5, and displays them on the conversion candidate display unit θ. Output and display every fixed number in association with the display number. In this embodiment, the number is five.

[Conversion candidate selection column 08: Selection instruction stage When the operator selects a specific candidate from among the conversion candidates output and displayed on the conversion candidate display section 6, the selection instruction input section 8 associates the conversion candidates with the candidates. Instruct the displayed number by pressing the number key.

609: Confirmed character string recording stage The conversion candidate selection unit 7 learns the conversion candidate word string corresponding to the numerical code input from the selection instruction input unit 8, and recognizes the words constituting it in the conversion candidate temporary storage unit 5. do. Thereafter, the notation of the conversion candidate is stored in the text storage section 12. The stored notation is displayed on the text display section 13 together with the previously input character string.

In the conventional character input device configured as described above and the character input method consisting of the steps described below,
− As long as there were related words in the units of text input, it was possible to convert them correctly. For example, "convenience" and "planning" are registered in the related word dictionary as related words. See Figure 7(d). At this time, the case where the reading ``Measure 1 Bengi'' is input and converted is inferior.

See Figure 4(a). Then, the kana-kanji conversion unit 3 creates a word with the pronunciation of ``bengi'' and a word with the pronunciation of ``hakaru'' as candidates in the conversion candidate temporary storage unit 5-.

See Figure 7 (1)). Next, the related word search unit 10 searches the word on the conversion candidate temporary storage unit 5 at this time in the related word dictionary 9.
If you search for 2, there is a pair of ``convenience'' and ``planning'', so is it the output priority of ``planning'' from among the homophones of ``measurement'', ``to measure'', ``measure'', and ``plan''? The first candidate was the conversion result "to aim for." See Figure 7(C).

Problems to be Solved by the Invention However, in the conventional character input device and method as explained above, since the reading is inputted by dividing it arbitrarily according to the operator's will, one of the characters between the divisions is input. A case may occur in which a related word pair does not exist in a conversion unit. In such a case, there is a problem that conversion using related word information cannot be performed. For example, if you enter "hengiwo" and "measure" as separate input units, after confirming "convenience", you will convert the "measure" part independently, ■I am unable to select the correct ``plan'' from homophones such as ``ru, plan, plan, measure, measure.''

Furthermore, when inserting characters into a fixed text, there is a problem in that conversion cannot be performed using the relationship with words in the fixed text.

For example, if you want to insert ``Please help me'' immediately after ``I would like to ask for your convenience'' in a sentence that has already been determined, ``I would like to ask for your convenience,'' the homophone for ``Measure'' (pronounced ``Measure''), ``Measure'', ”, “Plan”, “Measure”, “Weigh”
, could not suggest the correct word as the first choice from ``mote''.

SUMMARY OF THE INVENTION In view of the problems of the conventional character input devices, it is an object of the present invention to provide a character input device that can effectively utilize related word information and input characters efficiently in any situation.

Means for Solving the Problems The first invention provides a kana-kanji conversion unit that prioritizes and outputs conversion candidates for a kanji-kana mixed sequence corresponding to a Japanese pronunciation sequence, and a kana-kanji conversion unit that temporarily outputs conversion candidates output by the kana-kanji conversion unit. a conversion candidate temporary storage section for storing, a related word dictionary for storing word pairs that co-occur in sentences, a conversion candidate selection section for selecting candidates from the conversion candidate temporary storage section, and a conversion candidate selection section for selecting candidates from the conversion candidate temporary storage section; searching the related word dictionary to determine whether or not the word in the confirmed word temporary memory and the word in the conversion candidate temporary memory are in a related word pair relationship; This character input device is provided with a related word search section that determines the related word, and a related word priority change section that increases the output priority of the corresponding conversion candidate in the conversion candidate temporary storage section when the related word is related.

A second invention provides a conversion candidate generation step for generating an interpretation of a kanji-kana-mixed string corresponding to a reading in an information processing device having a related word dictionary that stores word pairs that co-occur in a sentence and a fixed word temporary storage unit. a related word search step of determining the relationship between the conversion candidate word and the word in the confirmed word temporary storage unit; and a word determined to be related among the conversion candidate word and the word in the confirmed word temporary storage unit. This character input method includes an output priority changing step of increasing the output priority of the word, and a confirmed word recording step of storing the finally selected candidate word in the confirmed word temporary storage unit.

A third invention is a patent claim that includes an input position changing section that changes an input position in a sentence, and an initializing section that initializes a fixed word temporary storage section when the input position is changed by the input position changing section. ■6 described in scope 1: Character input device.

A fourth invention is a character input method according to claim 2, which includes an input position changing step and an initializing step of initializing a fixed word temporary storage unit as the input position is changed.

A fifth invention provides a kana-kanji conversion unit that prioritizes and outputs conversion candidates for a kanji-kana mixed sequence corresponding to a Japanese pronunciation sequence, and a conversion candidate temporary storage unit that temporarily stores conversion candidates output by the kana-kanji conversion unit. a related word dictionary that stores pairs of words co-occurring in sentences; a conversion candidate selection section that selects candidates from the conversion candidate temporary storage section; a sentence storage section that stores at least sentence notation; and the sentence storage. an input position changing unit that changes the input position of a character in the text; a nearby word extraction unit that extracts words before and after the input position from the text storage unit; and a confirmed word that temporarily stores the words extracted by the nearby word extraction unit. a temporary storage unit, and a related word search unit that searches the related word dictionary to determine whether or not the word in the confirmed word temporary storage unit and the word in the conversion candidate temporary storage unit are in a related word pair relationship. , is a character input device including a related word priority order change unit that increases the output priority of a corresponding conversion candidate in a conversion candidate temporary storage unit when the word is related.

A sixth invention provides an information processing apparatus having a related word dictionary for storing pairs of words co-occurring in a sentence, a fixed word temporary storage section, and a sentence storage section for storing sentences, in which a combination of kanji and kana characters corresponding to a reading is arranged. a conversion candidate generation step of generating an interpretation; a related word search step of determining a relationship between a conversion candidate word and a word in the confirmed word temporary storage; and a conversion candidate and a word in the confirmed word temporary storage section. The method includes an output priority change step of raising the output priority of words determined to be related, and a nearby word registration step of extracting words before and after the character input position in the text storage section and storing them in the confirmed word temporary storage section. This is a character input method.

According to the above-described configuration, the present invention has a fixed word temporary storage section that stores several words that have already been entered into a fixed word, and it is possible to check whether or not the words therein are related to the words in the conversion candidate temporary storage section. Determination is made using the word search section. If relevant, increase the output priority of relevant candidate words.

Furthermore, if the input position is intentionally changed during character input, it is assumed that the input word series has lost its continuity, and the fixed word temporary storage section is initialized.

Furthermore, if the input position is intentionally changed during character input, the characters before and after the new character input position in the text storage section
The 11 words are copied to the confirmed word temporary storage section, and the related word search section determines whether there is a relationship between the words therein and the words in the conversion candidate temporary storage section. If they are related, the output priority of the related candidate words is set.

Embodiment 1 FIG. 1 is a block diagram showing the structure of a character input device according to an embodiment of the present invention. 1.2 of this configuration diagram.
3.4.5.6.7.8.9.11.12.13 and 14 are the same as the components with the same numbers in the conventional example shown in FIG.

Reference numeral 15 denotes a confirmed word temporary storage section, which stores up to five candidate words explicitly selected by the conversion candidate selection section 7.

The confirmed word temporary storage unit 15 is a FIFO storage type storage device, and when the sixth word is registered, the oldest word is automatically discharged.

Reference numeral 16 denotes a related word search unit, in addition to the conversion candidate words in the conversion candidate temporary storage unit 5''2, the confirmed words in the confirmed word temporary storage unit 15 are searched in the related word dictionary 9''2, and the conversion candidate words and confirmed words are searched. Related words are extracted from the confirmed words in the temporary word storage unit 15.

Reference numeral 17 denotes an initialization section, which initializes the confirmed word-time storage section 15 when the input position change section 14 changes the input position.

Regarding the character input device of this embodiment configured as shown in 2 below, its operations are divided into [reading ability], [input position change], [conversion candidate creation], and [conversion candidate selection] as shown in Figure 2. This will be explained using a flowchart.

[Reading box 101: Reading input stage The operator reads the reading of the character that he or she intends to input in the kana input section 1.
'' by pressing the corresponding Japanese syllabary keys.

Each time one key is pressed, the reading code corresponding to each key is sequentially additionally stored in the input character string temporary storage section 2. As soon as the characters are stored, the system waits for input again, and when a reading is input again, the reading input this time is concatenated with the reading input previously, and the continuation of the input reading string is stored in the input character string temporary storage unit. It is formed.

[Input position change unit 102: Input position change stage When the operator presses the cursor key on the input position change unit 14 with the intention of changing the input position in the text, the conversion candidate selection unit 7 selects the input position change unit 14. Confirmation processing is performed in response to the instruction using the currently displayed candidates. Thereafter, after changing the character input position in the text storage section 12''2, the change is displayed on the conversion candidate display section 6.

103: Initialization stage When the operator instructs the input position change unit 14 to change the input position, the signal is sent to the initialization unit 1.
7 is output. The initialization unit 17 initializes the confirmed word temporary storage unit 15 at that timing and clears the contents.

[Conversion candidate creation step 104: Conversion instruction stage When the operator presses the "conversion key" on the conversion instruction input section 4 to instruct the start of the process, the conversion instruction input section 4 receives the corresponding "conversion instruction signal." output to the kanji conversion unit 3. The kana-kanji conversion unit 3 starts conversion processing for the input reading in the input character string temporary storage unit 2 at that time.

105: Conversion candidate generation stage The Kana-Kanji converter 3 searches for a word that matches the beginning of the kana reading sequence in the input character string temporary storage unit 2 in the word dictionary held within the Kana-Kanji converter 3 and interprets it. Possible conversion candidate word strings are stored in the conversion candidate temporary storage section 5. Interpretable means that the input pronunciation is applied to the word and it is composed of an independent word and an attached word.
This refers to cases in which a phrase matches as a repetition of a phrase.

At this time, if multiple clauses are possible, interpret the first two clauses at a time and give priority to the one with the longest reading. The order of priority among homophones is the same as the order of storage in the word dictionary.

106: Related word search stage When the conversion candidate w word string is stored in the conversion candidate temporary storage unit 5, the related word search unit 16 converts the words stored in the conversion candidate temporary storage unit 5 and the confirmed word temporary storage unit 15 into a related word dictionary. Search within 49.

If the corresponding word does not exist, control is transferred to 108. If it exists, it is further checked whether the paired word of the searched related word dictionary record also exists in the conversion candidate temporary storage section 5 and the confirmed word temporary storage section 15. If it exists, the process proceeds to the output priority change step 107, and if not, the control moves to the conversion candidate display step 108.

107: Output priority change stage conversion candidate--time storage unit 5 and confirmed word temporary storage unit 15
If a related word string is stored in , the output rank of the corresponding word in the fixed word temporary storage unit 5 is set to 1 among homophones.

When there are multiple related word pairs, the shorter the distance between the words forming the related word pair, the higher the priority is given, and if the words are the same, processing is applied in the order of input.

108: Conversion candidate display stage The conversion candidate selection unit 7 creates conversion candidates by combining words with lower output priority of the conversion candidate words stored in the conversion candidate temporary storage unit 5, and displays the conversion candidate display unit 6-42. Output and display every fixed number in association with the display number. In this embodiment, the number is five.

[Conversion candidate selection] 109: Selection instruction step When the operator selects a specific candidate from among the conversion candidates output and displayed on the conversion candidate display section 6, the selection instruction input section 8 selects a candidate and a correspondence. Instruct the displayed number by pressing the number key.

110: Confirmed character string recording stage The conversion candidate selection unit 7 learns the conversion candidate word string corresponding to the numerical code input from the selection instruction input unit 8, and converts it into (7
4 are recognized in the conversion candidate temporary storage section 5''2.

Thereafter, the notation of the conversion candidate is stored in the text storage section 12.

The stored notation is displayed on the text display section 13 together with the previously input character string.

111: Confirmed! The 11-word recording stage conversion candidate selection section 7 stores the selected conversion candidate temporary storage section 5.
The second conversion candidate word is stored in the confirmed word temporary storage section 15. The confirmed word temporary storage unit 15 can store up to 5 words.
Since the number of registered words is limited to six, when registering the sixth word, the oldest registered word is removed immediately before that.

As described in 2 below, according to this embodiment, the fixed word temporary storage section has a FIFO storage structure, and old words that have become less relevant are automatically removed from the fixed word temporary storage section, thereby improving conversion accuracy. ”

Embodiment 2 FIG. 3 is a block diagram showing the configuration of a character input device in an embodiment of the present invention. 1.2 of this configuration diagram.
3.4.5.6.7.8.9.11.12.13 and 14 are the same as the components with the same numbers in the conventional example shown in FIG.

Reference numeral 15 denotes a fixed word temporary storage section, which stores up to five words extracted by a nearby word extraction section 18, which will be described later. The confirmed word temporary storage unit 15 is a FIFO storage type storage device, and when the sixth word is registered, the oldest word is automatically discharged.

16 is a related word search section, and a conversion candidate temporary storage section 5''-
In addition to the conversion candidate words, the determined fllL words in the determined word temporary storage section 15 are searched on the related word dictionary 9, and the words that are related between the conversion candidate words and the determined word temporary storage section 15 are searched. Extract.

17 is an initialization section, which initializes the fixed word temporary storage section 15 when the input position change section 14 changes the input position.

Reference numeral 18 denotes a neighboring word extraction section, which extracts words before and after the current input position in the text storage section 12.

Regarding the character input device of this embodiment configured as shown in 2 below, its operation will be described below with [reading power] and [input position change].
, [Conversion candidate creation], and [Conversion candidate selection] will be explained using the flowchart of FIG.

[Reading box 301: Reading input stage The operator reads the reading of the character intended to be input into the kana input section 1.
” - are given in order by pressing the corresponding Japanese syllabary keys.

Each time one key is pressed, the reading code corresponding to each key is sequentially additionally stored in the input character string temporary storage section 2. As soon as a character is stored, it enters the input waiting state again, and when a reading is input again, the reading input this time is concatenated with the reading input previously, and the continuation of the input string is stored in the input character string temporary storage section 2. formed into two.

[Input position change part 302: Input position change stage When the operator presses the second cursor key of the input position change part 14 with the intention of changing the input position in the text, the conversion candidate selection part 7 changes the input position change part 14. In response to the instructions of
Performs confirmation processing on the currently displayed candidate. after that,
After changing the character input position in the text storage section 12''2, the change is displayed on the conversion candidate display section 6''2.

303: Initialization stage When the operator instructs the input position changing unit 14 to change the character input position, the signal is output to the initialization unit 17. The initialization unit 17 initializes the confirmed word temporary storage unit 15 at that timing and clears the contents.

[Conversion candidate creation step 304: Conversion instruction stage When the operator presses the "conversion key" on the conversion instruction input section 4 to instruct the start of the process, the conversion instruction input section 4 converts the corresponding "conversion instruction signal" into kana-kanji. Output to the converter 3. The kana-kanji conversion unit 3 starts conversion processing for the input reading in the input character string temporary storage unit 2 at that time.

305: Conversion candidate generation stage The kana-kanji conversion unit 3 searches for a word that matches the beginning of the kana reading string in the input character string temporary storage unit 2 in the word dictionary held within the kana-kanji conversion unit 3 and can interpret it. The conversion candidate word string is stored in the conversion candidate temporary storage section 5. Interpretable means that the input pronunciation is applied to the word and it is composed of an independent word and an attached word.
This refers to cases in which a phrase matches as a repetition of a phrase.

At this time, if multiple clause breaks are possible, interpret the first two clauses at a time and give priority to the one that can be interpreted with the longest reading. The order of priority among homophones is the same as the order of storage in the word dictionary.

306: Related word search stage When the conversion candidate word string is stored in the conversion candidate temporary storage unit 5, the related word search unit 16 stores the words stored in the conversion candidate temporary storage unit 5 and the confirmed word temporary storage unit 15 in a related word dictionary. Search within 9.

If the corresponding word does not exist, control is transferred to 308. If it exists, it is further checked whether the paired word of the searched related word dictionary record also exists in the conversion candidate temporary storage section 5 and the confirmed word temporary storage section 15. If it exists, the process proceeds to the output priority change step 307, and if not, the control moves to the conversion candidate display step 308.

307: Output priority change stage When related word pairs are stored in the conversion candidate temporary storage unit 5 and the confirmed word temporary storage unit 15, the confirmed word temporary storage unit 1
Set the output rank of the corresponding word in 5 to 1 among homophones. When there are multiple related word pairs, the shorter the distance between the words forming the related word pair, the higher the priority is given, and if the words are the same, processing is applied in the order of input.

308: Conversion candidate display stage The conversion candidate selection unit 7 creates conversion candidates by combining words with lower output priority of the conversion candidate words stored in the conversion candidate temporary storage unit 5, and displays them in the conversion candidate display unit 6. Output and display every fixed number in association with the display number. In this embodiment, the number is five.

[Conversion candidate selection section 309: Selection instruction stage When the operator selects a specific candidate from among the conversion candidates output and displayed on the conversion candidate display section 6, the selection instruction input section 8 selects the candidate and the corresponding image. ” and the displayed number by pressing the number key.

310: Confirmed character string recording stage The conversion candidate selection section 7 learns the conversion candidate word string corresponding to the numerical code input from the selection instruction input section 8, and recognizes the words forming it on the conversion candidate temporary storage section 5. . Thereafter, the notation of the conversion candidate is stored in the text storage section. The stored notation is displayed on the text display section 13 together with the previously input character string.

311: Confirmed word recording stage The conversion candidate selection unit 7 stores the selected conversion candidate word on the conversion candidate temporary storage unit 5 in the confirmed word temporary storage unit 15.

Note that the number of stored words in the fixed word temporary storage section 15 is limited to a maximum of five, so when registering the sixth word, the oldest registered word is removed immediately before that.

According to this embodiment, the neighboring word extraction I mountain part 18 is activated at the first kana key after the manual position has been moved, so the processing load is lower than the processing required each time the key is moved. Since it is small, high responsiveness can be achieved.

Note that although the operation of the first embodiment has been described as updating the priority order of homophones, candidate words that have different word divisions may be targeted.

In this embodiment, two records are stored for one related word pair in the related word dictionary, but two records may be searched in both directions.

As described in detail, according to the present invention, related word information can be effectively used in any situation to perform efficient character input.

Specifically, it is possible to provide a character input device with a high conversion rate by applying related words even when converting kana-kanji into phrases. Further, when the input position is intentionally changed, the degree of conversion 1'i'f can be increased by excluding the confirmed word from the related word search. Furthermore, even when characters are input again at the changed input position, highly accurate character input takes into account the relationship with the word extracted from the confirmed text.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of a character input device according to a first embodiment of the present invention, FIG. 2 is a flow chart showing the procedure of a character input method according to a first embodiment of the present invention, and FIG. A block diagram of a character input device according to the second embodiment,
FIG. 4 is a flowchart 1-1 showing the steps of a character input method according to the second embodiment of the present invention. FIG. 5 is a block diagram of a conventional character input device, and FIG. FIG. 7 is an explanatory diagram for explaining the operation of a conventional character input device and method. 3φ...Kana-Kanji conversion unit, 5φ-・Conversion candidate temporary storage unit, 7.φφ conversion candidate selection unit, 9. Conflict/related word dictionary, 1.
1-...Related word priority order change section, 14.I! Kenjin power position change section, 15--・determined word-mouth, i memory section, 16...
Related word search section, 17@1111 initialization section, 18... Neighborhood word extraction section. Name of agent: Patent attorney Shigetaka Kurino Haka 1 name Diagram conversion candidate display area Related amount dictionary

Claims (6)

[Claims] (1) A kana-kanji conversion unit that prioritizes and outputs conversion candidates for a kanji-kana mixed sequence corresponding to a Japanese pronunciation sequence, a conversion candidate temporary storage unit that temporarily stores conversion candidates output by the kana-kanji conversion unit, and a sentence a related word dictionary that stores pairs of words that co-occur in the conversion candidate temporary storage unit; a conversion candidate selection unit that selects conversion candidates from the conversion candidate temporary storage unit; and a conversion candidate selection unit that temporarily stores conversion candidate words selected by the conversion candidate selection unit. a confirmed word temporary storage unit; and a related word search unit that searches the related word dictionary to determine whether or not the words in the confirmed word temporary storage unit and the words in the conversion candidate temporary storage unit are in a related word pair relationship. 1. A character input device comprising: a related word priority order change unit that increases the output priority of a corresponding conversion candidate in a conversion candidate temporary storage unit when the words are related. (2) In an information processing device that has a related word dictionary that stores word pairs that co-occur in a sentence and a fixed word temporary storage unit, a conversion candidate generation step that generates an interpretation of a kanji-kana mixed string corresponding to the reading; a related word search step of determining the relationship between the candidate word and the word in the confirmed word temporary storage unit; and a step of outputting priority for words that are determined to be related among the conversion candidates and the words in the confirmed word temporary storage unit. A character input method characterized by comprising an output priority change step of increasing the ranking, and a confirmed word recording step of storing finally selected candidate words in the confirmed word temporary storage section. (3) an input position changing unit that changes the input position in a sentence;
2. The character input device according to claim 1, further comprising an initializing section that initializes a fixed word temporary storage section when the input position is changed by the input position changing section. 4. The character input method according to claim 2, further comprising an input position changing step and an initialization step of initializing a fixed word temporary storage unit as the input position is changed. (5) a kana-kanji conversion unit that prioritizes and outputs conversion candidates for a kanji-kana mixed sequence corresponding to a Japanese pronunciation sequence; a conversion candidate temporary storage unit that temporarily stores conversion candidates output by the kana-kanji conversion unit; and a sentence a related word dictionary that stores word pairs co-occurring in the text; a conversion candidate selection unit that selects candidates from the conversion candidate temporary storage unit; a text storage unit that stores at least sentence notation; an input position changing unit that changes the input position of characters; a nearby word extraction unit that extracts words before and after the input position from the text storage unit; and a fixed word temporary storage unit that temporarily stores the words extracted by the nearby word extraction unit. and a related word search unit that searches the related word dictionary to determine whether or not the word in the confirmed word temporary storage unit and the word in the conversion candidate temporary storage unit are in a related word pair relationship, What is claimed is: 1. A character input device comprising: a related word priority order changing unit that increases the output priority of a corresponding conversion candidate in a conversion candidate temporary storage unit. (6) An information processing device that has a related word dictionary that stores word pairs that co-occur in a sentence, a fixed word temporary storage unit, and a sentence storage unit that stores sentences, generates an interpretation of a kanji-kana mixed string that corresponds to the reading. a related word search step that determines the relationship between the conversion candidate word and the word in the confirmed word temporary storage unit, and a related word search step that determines the relationship between the conversion candidate word and the word in the confirmed word temporary storage unit and an output priority change step of raising the output priority of the word determined to be, and a nearby word registration step of extracting words before and after the character input position in the text storage section and storing them in the confirmed word temporary storage section. Character input method.