JPS5933534A - Kana (japanese syllabary)-kanji (chinese character) converting and processing system - Google Patents

Abstract

PURPOSE:To improve the automatic converting rate without increasing the load of the operator, by storing different output order information depending on a word accompanied with declensional Kana and a word unaccompanied with said Kana in a Kanji word dictionary. CONSTITUTION:Nouns, verbs and adjectives or the like are stored in the Kanji word dictionary and further, identification information on whether or not accompanying a declensional Kana and the output order are stored further. For example, without a declensional Kana to a reading Kana ''Ko'', the characters are stored in the output order of ''Ko'' (child, powder)'' and with the declensional Kana, in the order of ''Ko (dense, inclusive)''. In forming a sentence, when the part converted into a Kanji accompanies the declensional Kana, the declensional key is depressed between the part converted into the Kanji and the declensional Kana to provide the identification code accompanied with the declensional Kana. Then, if there are homonyms, the Kanjis are outputted in the order with the declensional Kana. Then, the number of times depressing the next candidate key is saved and the load of the operator is releaved.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a kana-kanji conversion processing method used in Japanese word browsers, etc., and in particular specifies a portion to be converted as auxiliary information. t, l for kana-kanji conversion
A kana-kanji conversion processing method with an automatic conversion rate of 1-8 is used.

As conventional kana-kanji conversion processing J]1 (hereinafter also referred to as jlt to 1 conversion processing 1iItj) methods, there are known a Kanji part designation method and a conversion/no-conversion designation method.

The former specifies the part to be converted, for example, (I went to school). In this method, the four points are: strange man, strange man, beautiful man, good man, beautiful man, quiet. However, if the legal kana indicated by the mepelator changes, erroneous conversion is likely to occur. In order to prevent this, it is possible to register all the states 17 (1) in the single 11/'i word (11, but another problem arises in that the word 71↓ representation becomes extremely large.) 1~ Tfl ml'
-C1-line-
In addition to 1, ``stomach'', ``different'', ``will'', 1-wei'', ``here''
There is also %'ilf while an equal number of homophones 11jJ are output and the negative phase of the mepelator increases.

On the other hand, the second customer's d°shi・kan・fu(1 (conversion finger Shitoj method is, for example, watakushiroba H gakkou into* ``kuro (I go to school.)'' and -11, after converting `` To correct ``<'' to ``one'', enter JHWl in the final form, and after conversion, enter sen.

It changes to the conjugation 1 form of λ. According to this method, it is possible to prevent the increase in the homophone nji, but since it requires manual labor to change it to the final form and correction after conversion, there is a serious problem that the negative pressure of the mepelator becomes too large. ′
! There are 3.

The present invention (, ['') has been developed based on C& information, and its purpose is to solve the problems described in 1-1 of the conventional kana-kanji conversion processing method, and to reduce the capacity of word dictionaries. It is an object of the present invention to provide a kana-kanji conversion processing method that makes it possible to improve the automatic change J odor rate without increasing the burden on the operator.

One purpose of the present invention is to manually input an input sentence as a kana character string7.
In a kana-kanji conversion processing method that obtains a 1-li mixed sentence one by one, there is a means for specifying the part to be converted in the input kana character string, and a means for specifying the part to be converted into kanji, and if the part to be converted to kanji is accompanied by okurikana. In addition to providing a means for adding identification information for when it has been input and when it has not been input,
It's the first time to write a letter with an okuri kana.
By storing different output ranking information for each ft and ft, the conversion candidate single ff? Kana-kanji conversion process 1 characterized by changing the order in which + appears
This is achieved by one method.

In Kibatsu IIJ, ``A method of dividing the morphemes of (1j1() into the following three types (Nakano, !lrFri, ``11 UF
'+ Morphological Analysis J Iij Report Processing Vol. 20 No. 1o, 19
79 Ginsho).

(1) Among the auxiliary forms of Niga, a morpheme that is a component of a clause but has no substantive meaning in itself and is related to a syntactic function.

(2) Flj group: single i! It is the semantic core of /i, and is a morpheme that can also form single #4 by itself.

(3) #9: Letter: Single I! fi no compositionl1li1purple and sh”r f!7i Ring nido bond L, formal M, adds flavor, determines the nn goods tT, 1 i〆1: (grammatical character), but llj In Germany, it is not possible to form words with 2. r-shaped form v11 rice.

Nayo-1p7i (in-between) is divided into fQ1 as shown in Figure 1.

In addition, the parts of speech HA'1 are based on Hashimoto Grammar (see Gakuiku Grammar, Iwabuchi, Seki Ishi, Yamazaki, "New Edition I11 G1% Grammar", Shutsutsu Publishing, first edition, January 1972). A list of these is shown in Figure 2, a table of usage of terms is shown in Figure 3, and a bond diagram.

Hereinafter, embodiments of the adhesive according to the present invention will be explained in detail with reference to the drawings.

qc 5 Figure 11 A kana-kanji conversion processing device showing an embodiment of the present invention? / is a burr1 ivy diagram. In the figure I Ge
t Keyboard 17, 2G: [Input character buffer register,
5 is a control circuit, 4- is an output editor, 5 is a display device, 0 is a printer, 7 is a connection matrix table, 8 +: ] Hiragana-kana-kana correspondence table, and 9 is an output statement register. Also, 1
0 is independent n! The i word ``1 shi, 11 is a conjugation table, 12 is an attached'' 14 table, and 13 & J independent conjugation table.

Independent nor IWWl 0 is vr, tkiichisatsu kanji semi-closed dictionary (hereinafter simply referred to as [kanji section shojibi]), a part of which is shown in Figure 6. In addition to the IOA, there is a specialized language dictionary. 10B
It consists of three parts: , User Registration Language Dictionary 10C. Of these, 10 people had a pre-fI8 kanji single 11j4 dictionary.
In independent ir+, -・general nouns,・suhenma (・su゛variable nouns that can be accompanied by “suru”), verbs, adjectives “(d,
It memorizes things belonging to the adjectives #shrine, adjectives, adjectives, interjections, etc. Her memorization content includes 7 pronunciations (kanji only), 1 part of speech, conjugation, and identification information (*marked) whether or not it is accompanied by okigana, as well as a set of output rankings, etc.-C11
iI! They are arranged in the order of the mikana code.

Said specialized Wit i+/? In the evening 10B, only nouns (including the pseudonym system) are Hr! The contents of the memory are read as kana 1 table njl'5. .

The type (type of specialized Ill), output order, etc. are set as a set and arranged in the order of the reading kana code. In addition, the user registered word dictionary 10C also stores only nouns (including tro nouns), and the content is similar to the VIC memory content of the previous Fk3 J'F Gate Ri+Dictionvf l O13. ing.

Conjugation Table 1], part of its contents is shown in Figure 7. The conjugation IR' M is stored for each classification based on the conjugation form. However, the conjugation IR' M is stored in the ``Shibuya'' section, which is 91r with the * mark.

Line ma] number i- and i-o.

IMi':j'4121;1. Major classification table 12A
,! = It is divided into a minor classification table 12B, and the major classification table 12A is a formal noun, as part of its contents is shown in FIG.

Each auxiliary verb (formal verb) it/l l each auxiliary verb ni
It is classified by the 41st position of the 45th rib of the J l particle, and corresponds to Subdivision Table 12B by number. Minor classification table 12 BiJ whistle 9 As shown in Figure 9, each irl of formal nouns, conjugated forms of each layer of auxiliary verbs, conjugated forms of each layer of auxiliary verbs, and particles are classified in units of 11 [+]. It is something.

fg! Continuation column Table 7 is independent: (t) and Itl j', I true 1 It1 connection relationship, and the end as a clause ■ [In the table showing the ability, Independence 11! Classification table of i, i = J' M 17
4 classification table and vs lI! ;Duban°). FIG. 10 shows a part of the contents of the connection matrix table, and FIG. 11 shows the correspondence with the classification table.

Note that the control circuit δ is J23! This function is used to perform the 4S'A ability to control the relationship between a word 199 such as +3 and the table.

” - Book V, structured as described above! The operation of the example device will now be described.

(1) Input Input is performed using the keyboard device, l. An example of the character arrangement of the keyboard device 1 is shown in Fig. 12.

Honhoko board mounting M l? By combining the J and space keys and the special keys marked with ○, it is possible to specify whether or not to convert the character string that is subsequently input (', l).

The expression positions N and [I ( ) on the key tops of the character keys are not shown in FIG. 13 and are defined as Qll<.

Shift keys IA and LR3 are used in each mode.

It has upper scroll and lower scale νJ loquat ability, and shifts only while this is pressed) tJQ ability is activated/-J
:do. Shift key (hereinafter abbreviated as @ when explaining the 8-operation) IA and 113 are 1 exactly the same P
It is used to improve A-noh, and Torara may also be used.

The shift reverse key IC is a further development of the shift lock function, and after pressing this key, the shift type 1
and the no-shift state are all reversed.

Return to the original state by pressing the key again0) Yes0
Also, when you switch modes, it will automatically return.

Zuno”, shift reverse t1M f) i: l)
It is valid only within one mode and is not carried over to other modes. If you need to move to another mode and shift reverse, press the 1° mode change key, which is also used as a letter key.

When pressed at the same time as Pace Gear 11), the mode change function is activated. Mode change key position number, t,
Basically, it is desirable to arrange them symmetrically to the home position of each digit.

The following sets (classes) are defined for the mode change keys.

(]-) Alphanumeric mode; press the space key (hereinafter referred to as "■"), then enter alphanumeric mode - (hereinafter referred to as "○"). Same as evening/r brighter) can perform Noh. That is, (a) when inputting uppercase alphabetical letters and numbers for 1v (including some English letters RI'' and numbers), press the letter key displayed on the key top.

(b) When entering the lowercase letters of the alphabet, J: and English HC, press the corresponding letter key while holding down the Shift key], A or IB. However, only one station where the shift reverse key IC lamp is lit. In this case, the operations in (b) and (b) above are reversed.

In other words, when (n,), shift key IA or i-J,
1 Press B, hold down 7j and press the letter key, and then press the letter key at (b).

■) Kanakana mode: Press the above ■ and then use the Kanakana function. In this case, tJSJIs standard keyboard (C (3233)
It corresponds to the shift mode of Kana and Kana v18.

In other words, (a) When inputting kanakana and san-on symbols in the regular text, press the corresponding hiragana letter key.
For dakuon characters and semi-dakuon a characters, C is the dakuon symbol (IHI, ') zakan 'di',
Make leaf (1,'1ff) manually.

(b) Lowercase letters (a, i, tsu, ko, m, tsu, ya, yu1g
), When entering capital letters (wa, koto, ko) and kana symbols (1 shift key I A ttake 1 [3 while holding down), press the corresponding character key.

■) Hiragana mode: When pressing the hiragana mode key (hereinafter referred to as "■J and Fl13zu") after pressing the above ■
Hereinafter, this will be written as [○), ]. ) Subsequent operation 1/In the case of top level +J J I S standard keychain (CQ2)
33) corresponds to the shift mode of kana and kana symbols.

That is, (a,) When inputting the normal hiragana OJ:bi long sound 611, 1.1, its length: press the corresponding character key. Voiced characters.

Regarding the input of semi-voiced characters, it is the same as 'U Ll ').

(b) Lower case letters (a, i, u, e, o, tsu, ya, yu).

To enter capital letters (wo, ii, ゑ), and kana symbols, press the corresponding character key while holding down the Shift key I or 113.

(4) i1 character mode: Press +11 + mark ■' (from kanji mode l° ki basically l) From the character string after switching to this mode, press the next ■ and ○ (indicates any mode change key) ) and in this order; FI! The characters up to the next time the key is pressed are subject to kana-kanji conversion processing.

(1) For the simple noun 11), press the letter key with the 15 reading.

(b) Synthesis of name n, 11 It! , j GJnot; Separate the tree's closing 11jli and affix and press the character key for reading.

(c) 7tI N('ft1J Rlg, Akirayo 1f
iiili (verb) is 19; (also press the letter key for the slow reading of the character part.

) Okurikana mode: Press and hold the above ■ and select Okukana (
Hereafter, this is F@'doY114zu. ) subsequent Il? I.J.
When dealing with predicates using the upper button function mode, enter the okurikana part in the desired lFl form up to the 9th nori of zono 117i while switching to this mode.

(0) 'i'l' gate mode: Press Ill entry and then switch to specialized mode (hereinafter referred to as "(fi3 J and tIC
Talented. ) The subsequent action (4 Nomae [1)] is the same as the x-character mode 4; /i PI? 'N
! Kana-kanji conversion is performed using . I can't do it for a long time.

(7) User registered word mode: After pressing (S) above, press the user Jt, RNil mode key (hereinafter 1' (”))
Written as J. ) (hereinafter, this will be referred to as IOJ).
' (use conjugation)) Enter one input.

Above is a method of switching between each mode by pressing the space key (■) and then pressing the mode change key.
3- It goes without saying that you can use the /l/ key and mode change key, etc. in combination, or prepare a mode change key for ql.

The keyboard device t#1 includes the above keys, the -character kanji key 1iy, the 2-stroke key 'tF, the next candidate key IG, the update key], H, the delete key IJ +3, and the line feed key IK. etc. are being crotch IJ. The functions of these keys are explained below.

(1) - Character Kanji key IE When you press this, single n? You can enter ``11'' in the i-dictionary. ``-'' Enter the pronunciation for each character and press the next candidate key to convert it.

If you press the -character kanji key again, the -character kanji key mode will be touched or removed.

(2) Two-stroke key 11i' When this key is pressed, all characters can be input using a two-touch mnemonic. When the 2-stroke key is pressed again, the 2-stroke mode is canceled.

0) Blow A] Force supplement key IG The next candidate key is the key for requesting the next candidate in the kanji mode section V1, and is also the key for requesting the next candidate for the conversion request in the case of - character kanji 1,100.

(4) Update key 1.11 This key 4;1. lea character mode delimiter 111r * When selecting tart, )(≦the grandfather shown
lI [(is a key that requests the next candidate when the desired If!! is not found. The next candidate key is updated by I filIt, and the update key 1] is a key that requests the next candidate if the target If!! is not found.) As you can see, items d and d are skipped.

F delete key lJ This key deletes the 10th previous character "j" and the mode display, and can also cancel the "6-" code itself.

(0) i! fC line key], I (This key moves the character input from the next character key pressed to the next 7 screens i.

An example of input by the actual 1411 example device constructed as described above is shown below.

(J,) In general, 'sentence examples U I I J studies Iku he 1 EI 7. When creating the sentence ``ko'', the key operation is as follows, corresponding to the display tJ, J, -, N+, i■ - [phase] as shown in 1 and the following.

■It's me ■Akita ■Akita Ke ■I am Lji ■Let's attach a circle to me ■I [yo school Kasumi], ■I go to Gakkoden ■I go to school ■Do I go to school or kindergarten? [Phase] I'm the first in line to go to school (I'm 41) ■I go to school and O I went to school and started eating. [Ai] I am going to school first.

In addition, after -1- entry ■, even if you input as follows, the same result 5i will be obtained.
′! is lowered by 1.

This change in the display of the chorus is (Φ′I am circling around to school. 斡）'I go to schoolj ``It's 11.

Therefore, the 1st and 9th games can be easily completed by 4R, but there is a lot of internal processing in the 4 machines, so it takes some time.

(2) Same f, f Jj (Example 1 of 1r sentence containing i)
Attend the 5th celebration. ”, key 1)%
(J), and the display softens as shown below.

■Do the same ■Open vtrr nen1tTI Kr A fJrI D
, A. Open? ! ! +7,? hJ? tli 7.
Now, by pressing the next candidate key (as above), ◇) Opening a + ra + opening will result in 16 openings, 6 river ports, and revisions. Now that the desired "opening" has been obtained, move on to the next input. ■ At the school opening ■ School opening fence ■ School opening fence ■ School opening fence ■ Planting a field to celebrate the school's opening in 2018 ■ Attending the school opening celebration Circle (1) Attending the school opening celebration Attend the I3 celebration of the opening of the school.

As mentioned above, the same sound F? If there is 4, %l''S 1st place conversion i] force sleeve will be displayed in the normal position, and the 2nd place and lower conversion housode will all be displayed after the same sound 14i display (the above country).''2 For example: 15, we want to get the second place candidate, so if we press the next candidate key once, we will get the 11th place fi! 'h'J ``Open δjJ7 J'', force ζ will be deleted, and [゛Opening] will be displayed at that position. Along with this, 3. I! ! Shifted by i minutes.

[7, In this case, 1゛1f degrees [Kaigo-1 is converted to It is displayed, and in addition to the second place jiff (1 "opening 11 is displayed and below Ll
It will remain as before.

σ5) Synthesis t! 14 Examples of sentences containing '4- In the case of kanji (・I: karinorihi n, 'i, ? 9 goi, enter (su) again before the affix or 11 hanot t.

Taking as an example the J.S.A. that inputs "by celebration ceremony", the key operation is ``,'' and the display number changes as shown below and goes to 0.

6) Celebration Area Shikiten ■ Celebration Ceremony Mouth Also, when manually inputting ``1111+kai'', the key operation is , and the display changes to data as shown below and goes to 0.

■Area ■Great old area tccμm s△dai△daiG)Daisou■Great circle Undo■Daij11! Zoo ■ Dai j11 (Doori [Kai ■ Dai i1 (Doukai Inkai 6 times, 1 (1 shell, Kai, Kai, Borrow ■ Dairyoku Kaiguchi (4) Contains okuri kana'! J sentence example 1 If there is an okurikana in one of the characters &:l:, enter the second (seconds) before the okurigana part.

If you enter J without holding out, the key operation will be , and the display will change as shown below.

■ Same as enclosure ■Sashikuni Kusashi Δpoint 6) Scream Oka day pointing ■It's out of bounds ■Please send the same ■Open mouth ) Han il'f! Examples of sentences containing characters There are two ways to input "-12,300 yen".

The first key 11ψ is The display changes as shown below.

(Ri same place ■ place 1 η mouth U city / △True ■ - In the million range ■ - Manjiro picture 12. Load Δ boiled (smoked, - Man247■ - Man2 @ 1・υn 0 - Man2,000 [FM line, plug Δ selection △ , first, Cao ■ - 12,000 eyes 0 - 12,000 mouths α) - 12,03 area Japanese dialect, ugliness, praise, calculation, production dΦ - 12,03 round hi 0 - 1,200 Sanzono Biyaku (Φ - sword 2,300 garden (φ - 12,300 yen enclosure [Ai] - 12,300 yen mouth) - 12,300 yen mouth The key operations for inputting tα2 are (A) (A) stake), and the display changes as shown in the table of contents.

0 days 1 σ) 11 laps ■−・Ichiichi Ichigo g) -1 sip ■-Enclosed.

■-Mando CI) On the other hand 2 ■-Manji farm ■-Manji map 2■-Manji enclosure σ)-Manji enclosure @) -Manni thousand takushi 0-Manni thousand three garden 〇- The side of the M2,31 shrine 0 - M2,300 (?p-゛ M2 I・Mikuchibiyaku ■ - M2,300 0 - M2,304 [phase] - M2 1,300 yen enclosure, banquet, salt 0 - 12,300 yen den (0) Examples of sentences containing arithmetic numerals If you enter N5-/', rTl,, 21,567, -1, The key operation is as follows, and the display changes as follows.

■Qx 1.234. &07. -■T],, 234.5C37, - Flash Mata, "Showa 5
May 7, 15 FJ When inputting J, the key operation is , and the display changes as follows.

■Let's surround the Showa era ■Showa era Doujin short stories, funny stories, chants ■Showa period 5 ■Showa period 87 ■Showa 87 garden ■Showa b7 Ennen ■Showa convex 'γ year map ■Showa 0703 ■Showa! 57 Year 3 Year 0 August 1985 Seki ■ May 1988 IJ5 En (su ■ Rwa May 15, 1985 Mouth ('7) j Example of inputting the letter A [Itll'On, 'J' 600J When inputting , the key operation becomes , and the display changes to the following J: .

■[Cloud qxt t p Orjir t,
6 'C large [9, ■ It, 1 p o
tt 'r, o D @NTIE, when inputting ``I Ha to 11ohoo1.j, the key operation is , and the display changes as follows.

■Seki I△FZO△to△s o h, o om■ 工△
gOΔtOΔ1ohoo/, Seki (8) Example of a sentence containing katakana [Coding manual] When inputting, the key operation becomes ``, and the display changes as follows.

■Coding Manual■Coding Manual (1() Processing F4- #Fly 4 Figure is for this embodiment device): Kana-kaji conversion processing Jl [lJ essential not shown) 4-Chard. , 1:゛,
Based on this, the outline of the process will be explained.

α) Input processing (14A) 1) Receive the code from the board 1.

2) ¥11 of the code (class 1 by job).

3) Perform necessary processing depending on the type of code.

4.) Depending on the type of code, the necessary h processing unit (1-minute) is sent.

■) Number word dictionary search tough, b (1, 41, 1'), [)
In alphanumeric mode using the next candidate key, number 111ii
ff dictionary (Heading section: Arithmetic Suzu characters, Corresponding character section: Random numerals >'s: Search and convert f1 room sleeves.

(3) (i-character JII C dictionary # search processing (1/I C, l
4C') 1) Kanji +1j, nt in Kanji 1,000-do
4 Search the dictionary and make conversion candidates.

(4) Professional Nti 17th C! Search processing (i 4. I)
)] ) Sr'l M Mo F Nioi r:
qIr, the gate Pnti will be dismissed, and the change hall h1) will be set up.

◎) User registration 11c dictionary search process (1,4-rr:
)1) User M, user registration 1 in UP mode
! Search the Ml dictionary and create conversion candidates.

(0) Grammar) 1<1 processing (14F) l) Perform a connection test for okurigana.

('7) Item ``No. 1 Priority Processing (14・q)l) Conversion of parts of speech to be output preferentially (l? Do not put your sleeves forward.

2) Conversion value + of part of speech to be postponed? 'Ili is set to t←〇 (8) Output rank processing (14, H) l) Complement and rearrange the converted values based on the output rank information.

O) Display school jE processing (14J) 1) Display conversion candidates.

αO) Learning process (14, K) 1) Transformation U (switch the output order of nn, which was n-voked, and the previous supine correction.

2) A detailed flowchart of the input processing unit is shown in FIG. 15 (g), σ3). In Figure 5, P indicates code type check, Q indicates mode code separation and mode setting, It indicates processing mode identification, and S indicates kana-kanji conversion preparation. In addition, 11I
In the 15 yen figure, in addition to the abbreviation "M" for mode, alphanumeric characters, kanakana, and hiragana.

Kanji, delivery name, IP gate opening, and user registration mode are abbreviated as English M1 piece M, flat M, etc., respectively.

Below, following the process shown in the t1\15 diagram, nI? I will clarify.

(''5) Code type check (1)], ) If the input is the next candidate key code, pass it to the next candidate process (1
5A).

2) If the input is not the next candidate key code, check if there are any characters in the input sentence buffer register: r-tsuta (]
-5A) L/, (1'It+No Rere lJg7. Go to the next execution 1 of the Kasho Nitei treatment raft.

3) Check whether the input is a mode code or not (1
5c) If the mode code is not L, do not proceed to step 3filj where input characters are stored in the input sentence buffer register.

4) Codes other than the above are frequently used as mode codes and do not proceed to the next process.

(2) Mode code identification and mode Pi'5'rT (
Q) l) If the input is an okukana or name mode code, check the mode code register, if the previous mode was not kanji mode, go to error processing Un, if it is kanji mode, set the okukana handedness conversion processing flag. is set, and it is not passed to the pointer processing of the input statement buffer register (15I), 15
F.).

2) If the input is not in okurikana mode, the entire string in the input sentence buffer register is converted to the input sentence register (W).
and moves on to the next process (151), 15F).

3) If the input is in alphanumeric mode, select the input character = 1- or (some examples are shown in Figure 110).
Alphanumeric mode code storage rack β1! Pass it to (15G, IQ
H).

4) If the input is a kanakana mode code: Select ■ in the input character code table and pass it to the kanakana mode code storage process (15J + 15K).

5) If the input is user registration ffr+mode (": select ■ in the input character code table and enter the user), pass it to the mode code accumulation process (1511, 15M).

(15
G+ 1<5. J, 1. fiil, 15N).

7) If the human power is either hiragana mode code 1゛ or specialized jllll mode': 1-code, go to Jilt, the corresponding 1,000-code code storage.・If not, move on to the next step ill! (15P t 1
5 Q).

(3) Process] 111 mode identification (+'1. )]) -]t
'--Gooniced Code Regis, and the previous mode (previous mode) is English m character and kanakana.

In any of the hiragana modes, the output sentence 'v
! i (ml l) It does not pass to logic (lfilt, 1 convex'
1).

However, in alphanumeric mode, t, -) (for convex cases, Al
Are the letters in Aki numbers? -Ii or Zako r-Tsuku (1
,,5S), and if it is a number, it is passed to ν)(transfer processing).

2) The previous mode is specialized n-hamo-1゛Alike user registration 1
If 11b is in one of the 11 modes, (, 1, each corresponding dictionary #! is passed to the search process. If neither is the case, move to the next process ( ), 5 [J * 15 V).

(4-) Preparation for Kana-Kanji conversion (S) 1) Check the mode code register and select s 91 (new mode)? 1° (the mode before the previous mode in the character sequence (i11 mode)) For 1311 cases in IRI mode or Kanji mode chair 11, 4;t, corresponding 1) 5 1 word priority processing k
・After doing so, move on to the next process (], 8 W, ],
5X, 10Y).

2) If the new mode is Kanji mode and the previous mode) is in a mode other than Kanji mode, prefix priority processing is performed. Convex Z).

3) In addition, the above processing mode identification;
<You can also decide where to go. In FIG. 17, - indicates a mode other than kanji mode, and * indicates any mode.

Next, Kanji single 1m dictionary search and grammar l1l (Jijusuru It)
(See Figure ml-4 ni1> in Figure 18;
1! ! :I will explain.

(1) Nominal conversion 1) WI 杓 for which the send-kana flag is not set (1, perform processing centered on the nominal (188)).

2) Front f (1: It): Replacement input statement register (W)
The first O character in the middle is a vague single i! j j'r1' harm Jl,
1 bunch 1・J, exchange all matches 5i41i
J11i was released as J11i, and the next cabinet 1 or similar
I (l13'J') to move 1' (11313)
.

3) ,,1. -fi I searched for 1,1, -・to゛(
That's LH! '＋ too! If yes, suffix NY”
1LTj11 (10V) F (X l,, still -
A'rr The third-order case that did not occur is the angle [change j + y1 processing is not performed.

e) For the JL case where the JTJ↑1 centering is performed (IJ3A), the JTJ↑1-centered processing is performed (IJ3A).

2) Measure the length of the character string in 711 ↓ 01 seven points and r, the character string length in Jinriki Bunhojo Star 1) in 1), set it to k, and calculate the length from the stomach iM to the length in the city of Bo・r. Set to (
18C).

3) Exploitation in the conversion register (W),
(letters only)! /i トL/ te kanji simple) Ili Fi
? '1'
Contains one character fr +I:j All Kanji characters) 1t) Search the dictionary and transfer the entire result to /i'i' t% buffer (IESI)).

4) If there is no matching question in the above 1113 search, go to 11 without conversion.

(3) Word conversion, dictionary buffer processing 1) The beginning of the input sentence register (W) for conversion! Search the dictionary buffer for a character (181 bow).

2) If there is no match in the J2ft self search, add the next character of the pointer and return 1 (, 1)
Search the 3t buffer (181'').

3) L fti:', until a match is obtained.
2) Repeat.

4) The first letter in the dictionary is the digit iI in the dictionary.
1) (or, if the opening that matches the number of characters in the conversion input sentence register (W) is tr +; t, σ in the conversion input sentence register (W)) search the dictionary buffer for the first θ character 11j #I! Extract It as a conversion candidate and proceed to the next step.

5) In the above lIi search, -=Yi (Zurochinoga),
If there is no nit, proceed to non-conversion processing.

(4) Term conversion, conjugation processing 1)-Hi113 search, matching nlj term'
Enter the product IQ of only the product in the product name register <1. ) (18G).

2) I'i11 executive candidates? l! ill register (II
) (18G).

3) Set the remainder in the nl1 tail register (T) (18G).

The character string in the flR tail register ('T') is converted to each conjugation F for all parts of the part of speech name register (P).
Check with l/l JJ8 (181.1).

5) In the above search, if there is a match, l (If there is no Jfj, perform static 1N buffer processing 3;
Don't go back.

0) For the one part that matches, set it in the row -Tf・porridge (CN)?tfj''?L% register (CNIL) of the matrix table (18J).

) with bow nn processing 1) From one number register (CNIt), one number j'
Take out the financial name and check it to the CN pointer (1
8K).

2) Matched word M (Matakei・1 to αnii> ++D
In addition to 1 second of the character string in Executive 1/Jista (11), t? 74 Set in R section register (11) ( ],
8 L).

3) Matched i! IR (or Idama 1' 14 n!'
t) is subtracted from the beginning of the character string in the word tail register ('[') and set in the word tail register ('I') (181.).

4) Seven R6 processing r (+HH) g, +! i Jg club reply l
If a string remains in t('I') &'J next I
Moved to JL 311 (18M).

5) If there is no character string left in the n tail register (i'), the bank checks whether the string can be terminated for each connection matrix table, and if not, υ'' moves to the next 3 copies [1 (18N ).

0) Upper 8C: Check result is acceptable 41. &J:
, Conversion "f Do not proceed to supplementary processing (18N).

7) Check the connection "J'?!"j at the corresponding launch in the connection matrix table, and check the connection "J'?!"j that corresponds to the current column number.
Check the character string k in the 1113 tail register ('1'') using lj (181).

8) Accumulate all matching row numbers in the row Vli number register (CNR) 3111 ( 113 (,;!
, ＞.

9) - If not, check if the σ1 continuation can be connected from the next column ti, and when it reaches the end of the row, select C.
Check N pointer (1.8 liters).

10) If the CN pointer points to the point after the line number register (CNR), JJ, J
) Dictionary buffer after cleaning IJ record (・171+
=Return(], (3S).

11) J:, ML'! YK 414111 Report Rf, l-C7J case number, itcKn Return to the beginning of processing (188).

(Q) * * D rlff ”II 1141i&
3] Add the contents of IA (1814 minutes V) and add/f/f as follows. ) 1) If the number in the ward 91 part of the reading heading is 0 to 1 (
Jo, the number from the beginning of the character string in the reading heading is a number L) iB L,, convert the character string of the following into an output code, and then add it to the character string of the corresponding character 1 bar. do.

2) Perform storage processing with other contents (1, 8'l').

α) Conversion point storage ηII B 1) If the number in the delimiter part of the reading heading is other than 0, leave only the number indicated by the number at the beginning of the character string in the reading heading, and add 1 character string to the beginning of the character string in the reading heading. Convert the code to the output code and then append the corresponding character (closed string) to it.

2) After converting the character string in ll (T tail register ('r) into an output code, add it to the character string processed above.

3) Extract the dictionary contents and store them together with the contents other than blood in the 9th section of the 91st section of the dictionary (18U).

(8) Suffix processing l) Cut out 38 characters - C from the empty characters.

■) Non-conversion processing], ) Converting the character string (input character string) in the conversion input sentence register (ν■) to an output code and proceeding to display processing (1
8W).

Figure 19 shows the priority part of speech J7J! (1 in 14 causes)
Noyoj.

It shows the side of the body.

(1) If the Okukana flag is set, the 19th
Arrange (1)゛゛1　〇(2) in the order shown in the figure.
,) If the okurikana flag is not set, the 19th
The candidates are arranged in the order of ψ (which is completely different from the order shown in the figure).

(3) Furthermore, if the PK (prefix, general suffix, specialized suffix, auxiliary v&rliil') priority flag is set, the corresponding affix is arranged at the beginning.

In this case, the output order can be according to an algorithm such as that shown in section 17r''4 above.

Regarding output ranking processing, output ranking information j within the same part of speech is
This is dealt with by arranging it in IR. .

The following is a description of the two dictionary entries.

(1) Kanji word dictionary Figure 20 (a) shows the jjIt diagram.

1) Reading kana (reading heading) ((ze) Handles only hiragana, the code is I I 5CO226
The second byte (lower byte) of the 4th ward is used. This is designed so that the output hood, which will be described later, can be bent to ff)1ηtZ).

2) Notation (corresponding characters) Handles only kanji, call 1 -] j l S C622
(3, the 11th byte and the 2nd byte are both excreted!--('
H'1 hit (b6) is used as °°1゛1.

3) Part of speech/conjugation column Use JISC13220 code and use numbers, alphabets, and symbols. Gel about the ν-physical assignment of codes:
I￥ See Figure 19.

4) Output order (lock) Single■1) After extraction, display the display output order of homophone words.

do. In the learning process, this column will be changed to 1]).

5) Number of characters Express the number of characters to be output after the kana-kanji conversion process. Note that the heading delimiter indicates how many later characters to output out of the valid number of characters in the reading heading block, and the corresponding character section ('' indicates the effective number of characters in the corresponding character 11Vd).

Ω) Technical term dictionary A configuration diagram is shown in FIG. 20 (13).

1) Reading heading 19d 01 noki kanji 坩 f14t i! i'l
``It's the same as IF.

2) Character + 11 for pair rr The handling of all characters A11 (and the content 111ν are the same as the previous n11 kanji ゛:!1 rig).

3) 11 by fIl! +1 Specialty: 111-1) 11-specific, example sentences are general, place names 9 people's names 9
7B. and regulation 5 (','!-confirm. Code is JISC62
20 and 1 character table ↓I1. do.

4-) Output order It is exactly the same as the Kanji word dictionary mentioned above.

5) Number of output characters: 1194 Number of output characters from the fit character, il, j rL'i dictionary: 1
It is the same as the corresponding character part of 1.l except for "6134."

(3) User registration 11ij dictionary The configuration diagram is shown in Figure 20 (0).

1) Reading heading Jl #1 Hira (Ii name, alphabetical characters, handling of symbols and capacity are mentioned above. Kanji single) IM i! rl"iFJ.

2) Pair IJ1/λ character RF14 Except for handling all characters 01() and capacity, it is the same as the S-character monogram dictionary.

3) Type 1ff/1 Conforms to the technical term dictionary mentioned above.

4) Number of output characters 11・v It is the same as the technical language dictionary described above except for the capacity.

(4) Utilization costotail/tsuketsu 11μ table Figure 200 shows (U felling).

1) Other than the capacity of the reading heading field, it is the same as the above Kanji word dictionary. 02) Number of characters (the number of characters in the reading heading field is not shown. 0. 1
Indicates the row number or column number (see figure 0).

Next, an example of the coding system is shown below. By the way, is this part of the correspondence table between the output code and input character code of the residence board device? 1st
It is shown in Figure 6.

7i 11i of the input character code and output character code
Here is an example. The output character code is basically as described above,
J1. As specified in S Cf) 226, use bits b to b7 of 1 byte and 2 bytes as they are, and set bit 11 of each byte (b8) to °°1''.
shall be.

(1) Hiragana human power and ``'r'' code is set as lower bar r1・(ri), 28 bytes), upper byte cf official 1 byte) is A4-(
10100100) is added.

(Bll) Input character code Output character code
A 2 A4
A2 or A B
A4. A 1.1 is ACA, 4-
A C (2) The lower byte (2nd byte) is the kanakana human character code, and the upper byte (1st byte) is A5 (10100101).
) is added.

(Example) Input character code Output character two-door
A 2 A3A
2 forces A, HA, 5 A Bga
A, CA Convex AC (3) Change the highest byte (b8) of the input character code (both uppercase and lowercase) from °゛0°' to °“1”, then lower byte (first 2 bytes) and A3 (10100011) in the upper byte (tR1 byte).
Add.

(Example) Input character code Output character code (4) Since regular correspondence cannot be established for English symbols and kana symbols, for example, the second
A table as shown in Figure 1 is used.

The mode code is, for example, '9F' in the function character area of the JISC (3220B unit symbol).
Use the area of

1) Alphanumeric mode 81 2) Kanakana mode 82 3) Hiragana mode 83 4.) Okukana mode 84 Convex) Kanji mode 85 6) Specialized n11 mode 8゜7) User registration mode 87 8) Update mode 81ji 9) Next candidate mode 8F A list of the above code system from the code side is shown in Figure 22.

According to the present invention, in the kana-kanji conversion processing method that inputs an input sentence as a kana character string and sequentially obtains the corresponding kanji kana place and Shiribun, In addition to providing a means for specifying the part to be converted into Kanji and a means for adding identification information to indicate when the part to be converted to Kanji is input as accompanied by Okorigana and when it is not, the Kanji single N71 dictionary is provided with Okrigana. By storing different output ranking information for each case that is associated with the above and that that is not, the conversion candidate JK M is determined according to the presence or absence of the identification information.
Change the 111 force order of at W! This has the remarkable effect of realizing a kana-kanji conversion processing system that reduces the negative BB of the mepelator and improves the automatic conversion rate.

[Brief explanation of drawings]

Figure 1 - Figure 1 does not show the 11 classifications and conjugations; Figure 5 is a block diagram of a kana-kanji conversion processing device that is an embodiment of the present invention; Figures 0 to 11; The figure shows the contents of the dictionary, and Figure 12 shows an example of the character arrangement of the keyboard device.
FIG. 13 is a diagram without the keys; FIG. 14 is a flowchart showing an overview of the kana-kanji conversion process by the first example device of the present invention; and FIG. 15 (a). (I3) is a diagram showing the details of the main part, Figure 010 is a diagram showing a part of the input character code table, and Figure 17 is
Figure 18 is a flowchart showing details of process 33J1, without showing an example of a table for mode identification.
Figure 9 is a diagram showing a specific example of the priority part of speech section 3!11, txy
Figure 20 (A) - Diagram showing an example of the structure of a 0.0 Go dictionary, No. 21
The figure shows an example of output codes for symbols, and FIG. 22 is a list of code systems. 1: Keyboard device, 2: Input character buffer register, convex S control rji circuit, 4. + 7: L Power Edition Volume 1, 5 H Display Device, 6: Printer, 7: Connection Matrix Table, 8I Hiragana-Kana-Kana Correspondence Table, 9! Output sentence register, 10F independent language dictionary, 111 vocabulary usage table, 12: Attached! ! Fj table, 13:
Independent word buffer. Nogo ρ′ Ward Figure 4 Figure 7 Figure 8 (a) (b) 2A Figure 9

Claims (1)

[Claims] In a kana-kanji conversion processing method in which an input sentence is input as a kana character string and the corresponding sentences containing kanji and kana are sequentially converted (1), a means for specifying a part to be converted in the input kana character string; In addition to providing a means for distinguishing when the part to be converted to ``is'' input as accompanied by okrigana and when it is not, the kanji single nli dictionary also includes ``1v'' for 藺 with okrigana and ``so''. A kana-kanji conversion processing method characterized by changing the output 11[IY order of the conversion candidate single Hji according to the I1 and N/A of the previous rh2 identification information by storing different output order information for I and I.