CN1099647C - Alphabet inspection method utilizing pronunciation and wildcard - Google Patents

Abstract

The present invention relates to a spelling inspection method utilizing pronunciation and wildcards. Firstly, a phonetic alphabet converting principle is established according to an English pronouncing principle; then, character strings of phonetic alphabets are arranged, combined and replaced in groups according to the input sequence of the character strings of phonetic alphabets and the phonetic alphabet converting principle for generating a plurality of possible character strings of phonetic alphabets for outputting the character strings of phonetic alphabets; finally, the possible character strings of phonetic alphabets are filtered and outputted. In the present invention, pronouncing character strings, or/and wildcards can be simultaneously inputted, wherein the pronouncing character strings, the wildcards and the character strings of phonetic alphabets are separated for replacing phonetic alphabets.

Description

Utilize the spell check method of pronunciation and wild-character

The invention relates to a kind of spell check method, particularly relevant for a kind of spell check method of utilizing pronunciation and wild-character.

English spelling checks that (spell checking) is for an indispensable critical function in the application software of handling the English words data in the general computer, can judge on inspection whether the spelling of its English-word is wrong with the English-word that inputs in the computer.These application software of handling the English words data are except that Word (word processor), also comprise database (database) and trial balance (sreadsheet) or the like, use making the English manuscript or the individual character spelling in the English words data in the database of writing unlikely wrong.

At present general English spelling audit function can only be according to the monogram mode of whole individual character, and reaches simple spelling correction function, but but can't be according to close pronunciation to reach the function of spelling correction.For example, if user's desire is keyed in " mosquito " this English-word, but the user only remembers the pronunciation of this English-word, and forgets its spelling, therefore " maskito " is keyed in the memory body of computer.At this moment, general English spelling audit function can provide " masking " or " mascot " equiprobable correct spelling to select for the user on computer screen, but " mosquito " this word can not occur.This reason is that the monogram form of " mosquito " and " maskito " keep off.

Again for example, user's desire is keyed in " confidential " this English-word, but the user gathers " confidanshal " or " Kanfedential " when keying in computer again with general pronunciation, and general English spelling audit function all can't provide " confidential " to be possible correct spelling.

Owing to English show the language of literal for a kind of in the pronunciation mode, therefore, in fact, the user often has with general pronunciation and gathers into individual character when not remembering the spelling of English-word really, keys in the situation of computer again.Can't find out possible correct spelling with regard to pronunciation but the English spelling audit function of knowing as mentioned above, is general equal.

In addition, for English as for the user of a foreign language learning, English-word often have can't read the situation of standard very.If this user gathers into individual character with the approximate pronunciation of standard not too and keys in computer again, for example with " spend " spelling with " sbend ", with " student " spelling with " sdudent ", or with " school " spelling key entry computer with " sgool ", this moment is if the monogram form of spelling of keying in and correct spelling is approaching, for example " sbend " reaches " sdudent ", the English spelling audit function of then knowing still can reach " spend " " student " and classify possible correct spelling as, but when monogram form and correct spelling are near than water, for example " sgool ", the English spelling audit function of then knowing generally can provide " spool " or " stool " to be possible correct spelling, but can not provide " school " to be possible correct spelling.

Moreover, also have a kind of method of wild-character inquiry in the general spelling-checker, but it all is the inspection principles at spelling.Therefore, similar above-mentioned situation if the user keys in the English alphabet of a mistake, then can't find the English words that will inquire about forever.For example, c among the ci and the s among the si read " s ", if still the user keys in res*t, just can't find this word of receipt by the wild-character querying method of traditional spelling-checker.

In addition, when the user only knows the pronunciation of inquiry word, but for the spelling of want inquiry word and when uncertain, then traditional spelling-checker does not provide this kind to come the function of searching English word by phonetic symbol.

In above-mentioned background of invention, traditional many shortcomings that produced, fundamental purpose of the present invention is to provide a kind of spell check method of utilizing pronunciation and wild-character, its English-word with input judges on inspection whether the spelling of this English-word is wrong, and provide all possible correct spellings according to the pronunciation mode, and can phonetic symbol or the input mode of wild-character, to reach the purpose of spell check.

According to above-described purpose, the invention provides a kind of English spelling inspection method, can find out the English spelling inspection of possible correct spelling according to pronunciation.The software engineering method of this English spelling inspection can be with the English-word that inputs in the dual code mode in the computer apparatus, whether the spelling of judging this English-word on inspection is wrong, and, find out the close all possible correct spelling of pronunciation according to the pronunciation mode of the English-word of the spelling error of being imported.

According to above-described purpose, the invention provides a kind of spell check method of utilizing pronunciation and wild-character.At first, according to English pronunciation principle, set up a phonetic symbol conversion principle.Then, according to the input sequencing of phonetic symbol word string, and according to the phonetic symbol conversion principle, the permutation and combination of the phonetic symbol word string being done group replaces, to produce a plurality of possible phonetic symbol word strings outputs.At last, then filter possible phonetic symbol word string output.In the present invention, can import the English word string of pronunciation simultaneously, phonetic symbol word string, or wild-character, wherein, pronunciation word string, wild-character and phonetic symbol word string are separately to do the processing that phonetic symbol is replaced.

In embodiments of the present invention, filtration wherein comprises the following step at least.At first, with the phonetic symbol filtration of the consecutive identical repetition in the output of phonetic symbol word string, up to surplus next.If the k in the output of phonetic symbol word string or the last phonetic symbol of g phonetic symbol are n, then n makes η into; If the last phonetic symbol of the ∫ in the output of phonetic symbol word string is s, then s ∫ replaces into ∫, if the last phonetic symbol of ∫ is k, then k ∫ replaces into t ∫.Then, for each phonetic symbol word string output, seek suitable individual character, and each phonetic symbol word string in the phonetic symbol word string output is derivatized to similar phonetic symbol word string again, to seek suitable individual character, be used to avoid some consonant mute, and the mixing sound is difficult for distinguishing a tin leakage sound that is caused.

The present invention carries out English spell check by utilizing pronunciation and wild-character, can improve the possibility that occurs to the English of the English similar correct spelling of wrongly writing greatly, and make the user can carry out the input of the English that will check, accelerated to find the speed of the English of correct spelling by number of ways.

With reference to the accompanying drawings embodiments of the invention are described in detail below.

Fig. 1 illustrates the systems function diagram of the embodiment of the invention,

Fig. 2 shows the process flow diagram of the embodiment of the invention.

Fig. 1 has shown the systems function diagram of the embodiment of the invention.In the present embodiment, English spelling inspection 10 is the pronunciation principles according to english language, and cooperates English phonetic symbol (kenyon﹠amp for example; Knott, kk phonetic symbol) and wild-character inquiry (wildcard search) to reach more comprehensive and more to meet the purpose of user's demand.In the present embodiment, comprised pronunciation input inquiry 100 at least, class phonetic symbol input inquiry 102, pronunciation cooperate assonance target input inquiry 104, and pronunciation, assonance standard configuration close the input inquiry 106 of wild-character.Wherein, pronunciation input inquiry 100 parts are for person skilled in the art scholar, and the queried for items of Fig. 1 can be done other combinations of equivalence and revise, and can not break away from the suggested spirit of the present invention, all should be included in the following claim.

Fig. 2 has shown the process flow diagram of the embodiment of the invention.At first, input one inquiry word string in step 20, this word string can partly comprise English spelling, and part comprises wild-character, or part comprises the class phonetic symbol.For example, user's desire inquiry abbreviation word string, then it may import abiviA ∫ *, and wherein abivi is an English spelling, and A ∫ is the class phonetic symbol, and * is a wild-character.If the word string of being imported contains wild-character, then in step 22, wild-character made into the wildcard phonetic symbol; If the input word string also includes English spelling, then in step 24, will spell part and replace to the class phonetic symbol.At last, again with the English spelling principle, in step 26, do pronunciation inspection and inquiry.

The detailed step of pronunciation input inquiry 100 is as follows among the figure l.The method of English spelling inspection is according to KK (Kenyon﹠amp; Knott) phonetic symbol is the basis, and all individual characters are converted to possible phonetic symbol word string, utilizes each phonetic symbol word string to remove to seek the individual character that meets this phonetic symbol word string again.Though use the KK phonetic symbol in the present embodiment, yet other phonetic symbols also can be equally applicable to the present invention.

At first, analyze the individual character pronunciation character, this program comprises five steps, as beneath detailed explanation.

Step 1

(1) judges at first whether it is consonant entirely, if then can't correct;

(2) remove non-English and symbol "-" character in addition;

(3) consonant of the consecutive identical repetition of filtration makes it become single consonant.For example, convert apple to aple, convert suggest to sugest.But the special case of following table one is not then filtered.

Table one

cce hh oww cci tth sse ccy surre ss cch aww nn lly eww -ess

Step 2

(1), is divided into 26 groups with A to Z, as table two the word string of handling in the step 1

A is to shown in table two Z.Wherein special conventional letter is described as follows:

A:_, I ... Deng vowel

O：oI， I

R：

V: vowel

C: consonant

E: the breakpoint word string (as-ment ,-less ,-ly ,-ness ..., or the like)

X: null character (NUL) (NULL)

P: preceding have a character

N: after character (can be null character (NUL)) is arranged

According to word string put in order and according to table two A to the resulting replacement of table two Z, do group replacement and make and become following structure:

A (B, C, D) E (F, G) wherein, A, B, C, D, E, F, G are respectively sub-word string (SUBSTRING), it can be null character (NUL) (NULL CHARACTER), the sub-word string representative in the parantheses is interchangeable.

(2) (D) (F G) converts the full word string array shown in the following table three (FULL ATRING ARRAY) to E again for B, C above-mentioned A.Table three

ABEF ABEG ACEF ACEG ADEF ADEG

In the present embodiment, the maximum length of each full word string is 30, and the end code of word string is 0x00.The structure that is the word string array is STRING_ARRAY () (31).But this maximum length is particular length not, and the user can define suitable length arbitrarily.

Step 3

Then, filter by the word string array that is produced in the step 2 with the following step.

(1) vowel of consecutive identical repetition is filtered up to surplus next.

(2) if the last phonetic symbol n of k or g phonetic symbol then makes n into η.

(3) the last phonetic symbol as if ∫ is s, then s+ ∫=∫; If the last phonetic symbol of ∫ is k,

k+∫＝t∫。

Step 4

Determine following situation-ed pronunciation.

(1) if to be positioned at word string last for word string " ed ", and if the phonetic symbol in front be p, k, s, ∫, t ∫, θ, f, then the phonetic symbol of ed correspondence changes t into.

(2) if the situation of aforementioned (1) is false, and last phonetic symbol is not when being not d or t, and then the pairing phonetic symbol of ed changes d into.

(3) if aforementioned (1) reaches the neither establishment of situation of (2), then the phonetic symbol of ed correspondence changes Id into.

Step 5

(1) removes null character (NUL) in the word string array.

(2) go for each phonetic symbol word string (phonetic string) and seek suitable individual character.

(3) each phonetic symbol word string in the word string array is derivatized to similar phonetic symbol word string again, removes to seek suitable individual character.Its principle of deriving is as follows:

Since second phonetic symbol, running into has continuous consonant person, once removes a consonant, forms a new phonetic symbol word string.For example, prounouce is pr_nauns through the conversion phonetic symbol, then can be p_nauns, pr_naus, or pr_naun according to its extension of principle of deriving.This reaches the mixing sound and is difficult for distinguishing a tin leakage sound that is caused with being intended to be avoided some consonant mute.

Then, unified phonetic symbol sign indicating number is handled the similar problem of vowel, and the normalization of part phonetic symbol, with the corresponding phonetic symbol sign indicating number of being changed to the rule of table two Z by table two A.Comprised following (A) treatment principle in the present embodiment to (J).

(A) vowel is divided into following A, R, O number group, shown in following table four.Wherein, each corresponding code is to be code used in the present embodiment, is the convenience for routine processes.For person skilled in the art scholar, its other combinations that can do equivalence are revised, and can not break away from the suggested spirit of the present invention.

Table four

A group's (code 1)

R group's (code 2)

O group's (code 3)

Code 28

?I (SPECIAL-ED)

(B) consonant is encoded shown in following table five.Table five

(C) remove all stresses and secondary accent.

(D) if sound of ju (ju) is not at first syllable, and then the sound of j removes.

(E) if j before a phonetic symbol consonant, then the sound of j removes.

(F) if a phonetic symbol before the r Or , then will

Or Change into _.

(G) t ∫ (27) is if phonetic symbol t (22) before then removes t (22).

(H) if the phonetic symbol n before g or the k then changes n into η.

(I) If be not last phonetic symbol, then it is changed _ m into _ n, _ l respectively; If last phonetic symbol then reverts to m, n, l.Wherein Normalized mode is as follows.If alphabetical m, n, the l front have vowel (a, e, i, o, u, y), then its corresponding phonetic symbol be necessary for (m, _ n, _ l) or (

); If alphabetical m, n, the l front is not a vowel, then unavailable ( ) represent its phonetic symbol (but-sm except, that is sm → zm).

(J) remove the sound of continuous repetition, only keep one.

For the class phonetic symbol input inquiry 102 among Fig. 1, at first according to English pronunciation principle, set up a phonetic symbol and replace principle, as shown in Table 6, wherein, special conventional letter is described as follows:

A:_, I,

Deng vowel

O： I

R：

V: vowel

C: consonant

X: null character (NUL) (NULL)

P: preceding have a syllable

N: after syllable (can be null character (NUL)) is arranged

Then, being similar to the method for aforementioned pronunciation input inquiry, and putting in order and according to the resulting replacement of table six, do group's replacement and make it to become following structure according to the phonetic symbol string:

A(B，C，D)E(F，G)

Wherein, A, B, C, D, E, F, G are respectively the sub-word string of phonetic symbol (SUBSTRING), and it can be empty phonetic symbol character (NULL CHARACTER), and the sub-word string representative of the phonetic symbol in the parantheses is interchangeable.

With the method for similar aforementioned pronunciation input inquiry, (D) (F G) converts the phonetic symbol full word string array (FULL STRING ARRAY) shown in the following table seven to E again for B, C above-mentioned A again.

Table seven

ABEF ABEG ACEF ACEG ADEF ADEG

In present embodiment, the maximum length of each phonetic symbol full word string is 30, and the end code of word string is 0x00.But this maximum length is particular length not, and the user can define suitable length arbitrarily.

With the part principle of similar aforementioned pronunciation input query method, carry out the following step again.At first, filter by the phonetic symbol word string array that is produced in the preceding step with the following step.

(1) phonetic symbol of consecutive identical repetition is filtered up to surplus next.

(2) if the last phonetic symbol n of k or g phonetic symbol then makes n into η.

(3) if the last phonetic symbol of ∫ is s, s+ ∫=∫ then, if the last phonetic symbol of ∫ is k, k+ ∫=t ∫ then.

(4) remove to seek suitable individual character with each phonetic symbol word string.

(5) each phonetic symbol word string in the phonetic symbol word string array is derivatized to similar phonetic symbol word string again, removes to seek suitable individual character.Its principle of deriving is as follows:

Since second phonetic symbol, running into has continuous consonant person, once removes a consonant, forms a new phonetic symbol word string.For example, prounouce is pr_nauns through converting phonetic symbol to, then can be p_nauns, pr_naus, or pr_naun according to its extension of principle of deriving.This reaches the mixing sound and is difficult for distinguishing a tin leakage sound that is caused with being intended to be avoided some consonant mute.

Cooperate assonance target input inquiry 104 for the pronunciation among Fig. 1, in the middle of this kind inquiry, pronunciation part and class phonetic symbol partly are separate processes.For example, when the user in inquiry during abbreviation, it may import abiviA ∫ ione, wherein abivi and ione are English spelling, A ∫ is the class phonetic symbol, A ∫ and abivi and the phonetic symbol that ione changed out are separate processes.That is to say that pronunciation partly is to utilize table two A to handle to the conversion method of table two Z, the phonetic symbol part is then handled by the conversion method of table six.Phonetic symbol word string after combination is finished is launched according to aforementioned table seven method of deploying again.With top input example is example, does following explanation.At first, it is as follows to do conversion respectively for pronunciation and phonetic symbol part according to table two A respectively to table two Z and table seven:

abbrivi?→?AbrAvA

A →?A

∫ →(∫， ₃，d ₃)

ion →An

e →(null，A)

At last, be combined into AbrAvAA (∫, 3, d3) (null A), after launch filtering, can obtain following various possible phonetic symbol word string: AbrAvA ∫ An, AbrAvA ∫ AnA, AbrAvA to An _3An, AbrAvA ^d3An, AbrAvA ^d3AnA, ArAvA ∫ An, ArAvA ∫ AnA, ArAvA _3An, ArAvA3AnA, ArAvA ^d3An, ArAvA ^d3AnA, AbAvA ∫ An, AbAvA ∫ AnA, AbAvA _3An, AbAvA3AnA, AbAvA ^d3An, and AbAvA ^d3AnA.

Close the input inquiry 106 of wild-character for the pronunciation among Fig. 1, assonance standard configuration, in the present embodiment, be with * or? as wild-character, wherein * represents a phonetic symbol or a string phonetic symbol, and? represent single phonetic symbol.Speech querying method and general wild-character querying method are similar, but maximum difference is to be substituted is phonetic symbol but not English-word.The input inquiry 106 that pronunciation, assonance standard configuration are closed wild-character has following several array configurations again:

(1) pronunciation cooperates the wild-character inquiry.For example, the user imports buti*, then can find beaut by this querying method, beauteous, words such as beautiful.Again for example, when the user imports re? it then can find repeat by this querying method, resit, words such as receipt.

(2) phonetic symbol cooperates the wild-character inquiry.For example, the user imports kAn* ∫ Al, then can find words such as confidential by this querying method.

(3) pronunciation, assonance standard configuration are closed the input inquiry of wild-character.For example, the user imports konfAdAn*, then can find confidence by this querying method, confident, words such as Confidential.

Table two A:A group

Xa-: a → A XaX: a → A XaweE: awe → A XawC: if C1=' h ' a → A else aw → A Xaw: aw → (A, Aw, w) awing: awing → O η awi: awi → (AwA, O) awh: awh → (Ah, Ahw) awyX: awy → O CawC: aw → A aw: aw → (A, Aw) aux: au → A air: air → Ar areE: are → (Ar, ArA) ParE: ar → (Ar, R) ParC: ParedX: CalE: al → (Al, l) (a is not first syllable) Xay: ay → (A, Ai) else: a → A

Table two B:B group

Xb-: b → bA XbX: b → bA bone: bone → (bAn; BAnA) (back does not meet r) boreC: bore → (bAr, bArA) (C1 unequal to ' r ', ' d ') Xbh:bh → b Xbd:bd → d XbaleN:bal → (bAl, bAlA) (N1 unequal to ' r ') XbakeN:bak → (bAk, bAkA) (N1 unequal to ' r ') XbaneN:ban → (bAn, bAnA) (N1 unequal to ' r ')

XbareN ：bar →(bAr，bArA) (N1！＝‘r’) XbaksN ：bas →(bAs，bAsA) (N1！＝‘r’) Pb ：b →b else ：b →b

Table two C:C group

Xc- ：c →sA XcX ：c →sA PceE ：ce →s (P！＝‘-’) counter?N ：if?N1＝‘e’，cout?→kAnt： else →kAntR cunter clothes ：clothes →klAz cci ：cci →(ksA，t∫A) ccy ：ccy →(ksA，t∫A) cceN ：cce →(ksA，t∫A) (N1！＝‘r’) ceuse ：ceuse →sRz ceous ：ceous →∫_s cross ：cross →krAs PciN ：if P＝X， c→(∫，s) (N1！＝‘a’，‘e’，‘o’，‘u’) else c→(special?∫， s) PceN ：c →(s，∫) (N1！＝‘a’，‘o’) cN ：c →s coe ：c →(s，k) ck ：ck →k Xchol ：ch →(k，t∫) Xchem ：chem →(kAm，∫Am) chN ：chr →kr (N1！＝‘r’and N！＝E and N2！＝‘e’，N3！＝null) XchN ：chl →kl (N1！＝‘l’and?N！＝E) PcN ：ch →(k，t∫，∫) (N1！＝‘h’and N！＝E) ch ：ch →(t∫，∫，k) Xcze ：cze →t∫A

crew ：crew →krA Xcre ：cre →krA cre ：cre → (krA；kR,kAr) cae ：cae →sA XcuN ：c →kw (N1！=' i ', ' e ') XcN: c → NULL (N1!=' l ', ' r ', and N1=C) the last phonetic symbol of Pc: if is k. c → NULL else c → k else: c → k

Table two D:D group

Xd- ：d →dA XdX ：d →dA Xdh ：dh →d Xdj ：dj →d 3 disch ：disch?→dAst∫ Pdg ：dg →(d 3，dg) Pdia ：d →(d 3，d) dz ：dz →(d 3，dz) duo ：du →(d 3A，dA) PdulV ：du →(d 3A，dA) Pdu ：d →(d，d 3) XdN ：d →NULL (N1＝C and N1！＝‘r’，‘w’) PdC ：d →(d，dA) (C1！＝NULL) else ：d →d

Table two E:E group

Xe- ：e →A XeX ：e →A eer ：eer →(Ar，AR) ereN ：ere →Ar (N1！＝‘x’and?N＝E) Xelse ：else →Als ee ：ee →A

EN (N1='r ' and N unequal to E) Xew PewN (N1!=' h ') Pexu earN Xeu Peur essE eauxE CesE PeE PedE (e is not first syllable) CelE (e is not first syllable) CeC (e is not first syllable) else

: if N2='e ', N3='d ', N4=E else if N2='e ', N3 unequal to ' r ', N3=C else if N2='s ', N3=NULL else if N2 unequal to NCLL; N2=C else: ew: if P=C, N1=C else: ex: if N1=NULL else: eu: e: ess: eaux: if C=' t ' else: if e is first syllable else if P=' y ' the or P=C: ed: el: e: e

?er→(Ar，R) ?ere→(ArA，Ar，RA ?) ?er→(Ar，R) ?er→(R，Ar) ?er→(R，Ar，r) ?→jA ?ew→A ?ew→(A，Aw，w) ?→(Ak∫，Aks) ?ear→(Ar，R，AR) ?ear→(Ar，R) ?→jA ?→(NULL，A) ?→As ?→(A，Az) ?es→(Az，As，s) ?es→(Az，As，z，s) ?e→A ?e→(NULL，A) ?→(special?Id， ?Ad) ?→(Al，I) ?→(A，NULL) ?→A

Table two F:F group

Xf- ：f →Af XfX ：f →Af fore ：fore →(fAr，fArA) XfN ：f →NULL (N＝C?and N1！＝‘l’，‘r’，‘j’ ) else ：f →f

Table two G:G group

Xg- ：g →d 3A XgX ：g →d 3A

?Xgh ：gh →g ?gh ：gh →(f，gh，NULL) ?guard ：guard →(gArd，gRd) ?guerN ：guer →(gR，gAR) ?(N＝C?or?N＝E) ?guV ：g →(gw，d 3) ?gu ：g →(g，d 3) ?XgN ：g →(g，d 3) ?(N1！＝‘a’，‘o’，‘u’ ?) ?gV ：g →(d 3，g，3) ?XgC ：g →NULL ?(C1！＝l’，‘r’) ?PgC ：g →(g，gA) ?(C1！＝NULL) ?else ：g →g

Table two H:H group

?Xh- ：h →At∫ ?XhX ：h →At∫ ?PhN ：h →NULL ?(N＝NULL?or?N1＝‘-’) ?hourC ：hour →Ar ?half ：half →hAf ?Xheir ：heir →Ar ?XhonV ：hon →(hAn，An) ?-honV ：hon →(hAn，An ?hua ：hua →(wA，hA) ?else ：h →h

Table two I:I group

Xi- ：i →A XiX ：i →A Xisu ：isu →AsA issue ：issu →AsA irer ：ir →Ar ireC ：ire →(Ar，ArA) ire ：i →A

Iron: iron → (ArAn, ARn) ir: ir → (Ar, R) ier: ier → (Ar, AR, R) iard: iard → (jRd, jArd) vi: A and i synthesize → O (last phonetic symbol is A) else: i → A

Table two J:J group

?Xj ：j →d 3A ?XjX ：j →d 3A ?XjesV ：jes →(d 3Az，d 3As) ?Xj ：i →d 3?else ：j →(d 3，3，j)

Table two K:K group

Xk- ：k →kA XkX ：k →kA Xkn ：kn →n kh ：kh →kh XkC ：k →NULL (C！＝‘r’，‘v’，‘l’，‘w ’) PkC ：k →(k，kA) (C！＝NULL) else ：k →k

Table two L:L group

Xl-: l → Al XlX: l → Al Xlh: lh → l lery: le → (lA, Al, AlA) PleC: if P=' c ' le → (l, Al) (P! The and of='-' C1!=' r ' else le → (Al, l) and le is not first syllable) lj: lj → l

?else ：l →l

Table two M:M group

?Xm- ：m →Am ?XmX ：m →Am ?Xmision ：mi →mA ?XmishC ：mi →mA ?Xmis ：mis →(mAs，mAz) ?Xmh ：mh →m ?Xmn ：m →NULL ?Xmc ：mc →mAk ?Xmb ：m →Am ?else ：m →m

Table two N:N group

Xn- ：n →An XnX ：n →An Xng ：ng →Ang PngE ：if?P＝‘i’，‘o’，or?E＝NULL ng→η else ng→ηg nguN ：if?N2＝NULL ng→ηg (N＝NULL?or else ng→η N1＝‘d’，N2＝E) ngs ：ngs →ηz ng ：n →n niorX ：nior →(nR，nAR) news ：news →(nAz，nAs) Pnazi ：nazi →(nAtsA，nAzA) (P＝NULL or P＝‘-’) else ：n →n

Table two O:O group

Xo- ：o →A XoX ：o →A Xoe ：o →NULL Pour ：o →NULL

Oine: oin → (wAn, On) oyC: oy → O oy: oy → (O, Aj) oirX: oir → Or (o is first syllable) oir: oir → (wAr, Ar, R) owyX: owy → (O, AwA) owi: owl → (O, AwA, A) owh: owh → (Ah, Awh) CowC: ow → A ow: ow → (A, Aw, w) oar: oar → (Ar, R) oreE: ore → (Ar, ArA) or: or → (Ar, R) ourE: our → (R, Ar) ough: ough → (A, Af) oeyX: oey → O oeN: oe → (O, A) (N1! ＝‘r’) oi ：oi →O XoneE ：one →wAn Xone ：one → (wAn；An) oneE ：oue → (wAn；An) one ：o →A Xonce ：once →wAns XofE ：of → (Av,Af) else ：o →A

Table two P:P group

Xp- ：p →pA XpX ：p →pA Xphth ：phth →θ Rhood ：p →p phorn ：p →p phole ：p →p phouse ：p →p PphV ：ph (f，ph) ph ：ph →f Xpt ：p →NULL Xpn ：p →NULL

?Xpc if?C2＝C p→p ?(C1！＝‘f’，‘l’，‘r else p→NULL ?’) ?PpC ：p →(p，pA) ?(C1！＝NULL) ?else ：m →p

Table two Q:Q group

Xq- ：q →kA XqX ：q →kA quV ：qu →kw quC ：qu →kA else ：q →k

Table two R:R group

Xr- ：r →Ar XrX ：r →Ar Xrh ：rh →r Xrw ：rw →r CredX ：re → ( rA； R) Pr: r → NULL (P=' r ') CreE: re → (R; RA) (e is not first syllable, dawn C unequal to '-') else:r → r

Table two S:S group

Xs- ：s →As XsX ：s →As Xsb ：sb →(zb，sp) Xsd ：sd →st Xsg ：sg →sk Xsh ：sh →∫ sh ：sh →(∫，sh) Psch ：s →s (P＝‘e’) sch ：sch →(∫，sk) sc ：s →(s，sk)

seuse ：seuse →sRz XsugV ：s →(∫，s) ssur ：s →NULL Xsure ：sure →(∫Ar，sAr) sure ：sur →(∫R，∫Ar， 3R) Xsur ：sur →(sR，sAr) sur ：s →(s，∫， 3，z) suN ：suN1 →(swA， 3A，∫A) (N1＝‘a’，‘o’，‘i’and N1！＝‘e’，N2！＝‘r’) su ：s →(∫， 3, z) (the last phonetic symbol of S is not k) su: s → (∫, 3) (the last phonetic symbol of S is k) sier: s → (∫, 3, z) (the last phonetic symbol of S is not k) sier: s → (∫, 3) (the last phonetic symbol of s is k) sia: s → (s, ∫, 3, z) (the last phonetic symbol of S is not k) sia: s → (s, ∫, 3) (the last phonetic symbol of S is k) XsmE: s → s PsmE: s → (z; S) sm: sm → z Psion: sion → (∫ An3An) PsiE: si → (s, z, sA, zA) (E!=NULL and i be not first syllable PseE: the se → (E!=NULL and (s, z, sA, zA) e is not first syllable XsV: s → (s, z) sE: s → (s, z) sN: s → (s, z) (N1!=' a ', ' i ', ' o ', ' u ', ' y ') Ps: s → (s, z) (P1!=' a ', ' i ', ' o ', ' u ', ' y ') else: s → s

Table two T:T group

Xt- ：t →tA XtX ：t →tA Xtongue ：tongue →tAη Xtsw ：tsw →tsw Xts ：ts →(ts，z) Xth ：th →(θ，_，t) thN ：th →θ (N＝NULL or N1＝‘-’) th ：th →(θ，_，th) tz ：tz →ts tureE ：ture →(t∫R，t∫Ar，tAr) ture ：t →(t∫，t) tsch ：tsch →(t∫，t+∫) tch ：tch →(t∫，t) Ptu ：t →tAp Xtype ：type →t∫An Ption ：if?P＝‘s tion→t∫An else tion→∫An PtiN ：if?P＝‘s’ t→(t∫，t) (N1！＝‘a’，‘e’，‘?else t→(∫，t) o’，‘u’) PtC ：t (t，tA) (C1！＝NULL) else ：t →t

Table two U:U group

Xu- ：u →jA XuX ：u →jA XurC ：ur →R Xur ：u →jA Xunkn ：unkn →An Xuc ：if?C2＝C?or?C1＝‘n’，C2＝‘e’ u→(A，jA) else u→(jA，A) uoy ：uoy →O uor ：u →NULL ureE ：ure →(Ar，R，jR) urer ：ur →Ar

ures ：ur →Ar ure ：ur →(Ar，R，jA) ur ：ur →(R，Ar，jA) PueE ：ue →(NULL，A) (P＝‘g’) PuedE ：ue →(NULL，A) (P＝‘g’) else ：u →A

Table two V:V group

Xv- ：v →vA XvX ：v →vA vE ：v →(v，f) vior ：vior →(vR，vAr) viour ：viour →(vR，vAr) Xvice ：vice →(vAs，vAsA) XvC ：v →NULL (C1！＝‘l’，‘r’) else ：v →v

Table two W:W group

Xw- ：w →dAblA XwX ：w →dAblA wh ：wh →(wh，h，w) wr ：w →NULL ward ：ward →(wArd，wRd) walk ：walk →wAk else ：w →w

Table two X:X group

Xx- ：x →Aks XxX ：x →Aks XxC ：x →(Aks，krAs) Xx ：x →z xion ：xion →k∫An xious ：xious →k∫As xual ：xual →k∫Al else ：x →(ks，gz)

Table two Y:Y group

Xy- ：y →wA XyX ：y →wA XyC ：y →(A，j) Xy ：y →j yrC ：yr →(R，Ar) yV ：y →(j，A) else ：y →A

Table two Z:Z group

Xz-: z → zA XzX: z → zA Xzw: zw → tsw Xzinc: zinc → (zA η k, zAns) zue: zue → zwA Pzu: the last phonetic symbol of if z be z → (3，s) k， z→( 3, z) the last phonetic symbol of else Pzier: ifz be z → ( 3，s) k， z→( 3Z) else XzV: z → (z, s) (the last phonetic symbol of z is k) else: z → z to Pz: z → s

Table six

Word string w phonetic symbol XwV w (w after the replacement of phonetic symbol word string string wish replacement word string; Hw) wh wh (hw; W; H) wr w null j phonetic symbol Cj j null PjA j null h phonetic symbol Xhw hw (hw; W) PhN h null, (N=null or N1='-') m phonetic symbol Cm m, (m; Am) Xmh mh m Xmn m null Xmc mc mAk Xmb m Am n phonetic symbol Cn n (n; An) Xng ng Ang PngE ng η Png ng (ng; η) l phonetic symbol X1h lh l C1X l (l; Al) r phonetic symbol Rr Rr (Ar, R) Xrh rh r Xrw rw r f phonetic symbol XfN f null (N=C and N1 unequal to ' l ', ' r ') v phonetic symbol vE v (v, f) XvC v null (C1 unequal to ' l ', ' r ') θ phonetic symbol X θ θ (θ, _, t) θ θ (θ, _) _ phonetic symbol _ _ (_, θ, t) _ _ θ _ _ (_, θ) s phonetic symbol Xsb sb (zb, sp)

Xsd sd st Xsg sg sk Xsh sh ∫ sh sh (∫, sh) XsmE s s PsmE s (z, s) sm sm z XsV s (s, z) sE s (s, z) sN s (s, z) (N1=vowel) Ps s (s, z) (P!=vowel) z phonetic symbol XzV z (z, s) kz z s Xzw zw tsw ∫ phonetic symbol P ∫ ∫ (∫, d, d 3) ∫ ∫ (∫; T ∫) b phonetic symbol Xbh bh b Xbd bd d Pb b b PbC b (b, bA) (C1 unequal to null) p phonetic symbol PphV ph (f, ph) ph ph f Xpt p null Xpn p null XpC if C2=C P (C1 unequal to ' f ' p, ' l ', ' r ') else null p PpC p (p, pA) (C1 unequal to null) d phonetic symbol Xdh dh d Pdg dg (d3，dg) dz dz (d 3，dz) XdN d null (N1＝C and

=' r ', ' w ') PdC d (d, dA) (C1!=null) t phonetic symbol Xts ts (ts, z) Xth th (θ, _, t) thN th θ (N=null or N1='-') th th (θ, _, th) tz tz ts PtC t (t, tA) (C1!=nul l) g phonetic symbol Xgh gh g gh gh (f, gh, null) XgV g (g, d 3) gV g (d 3，g，3) XgC g null (C1 unequal to ' l ', ' r ') PgC g (g, gA) (C1!=null) k phonetic symbol Xkn kn n kh kh kh XkC k null (C unequal to ' r ', ' v ', ' l ', ' w ') PkC k (k, kA) (C1 unequal to null) з phonetic symbol P3 3 ( 3，d 3) (P！＝null) 3 3 ( 3，d 3，∫) d 3Phonetic symbol P d 3 d 3 (d 3，3) (P！＝null) d 3 d 3 (d 3，3, ∫) t ∫ phonetic symbol t ∫ t ∫ (t ∫, ∫)

The above is the preferred embodiments of the present invention only, is not in order to limit claim of the present invention; All other do not break away from equivalence change or the modification that disclosed spirit is finished, and all should be included in the protection domain of following patented claim.

Claims (9)

1. class phonetic symbol spell check method comprises the following step at least:

According to English pronunciation principle, set up a phonetic symbol conversion principle, be the basis promptly according to a kind of phonetic symbol, all words are converted to possible phonetic symbol word string;

According to the input sequencing of phonetic symbol word string, and according to this phonetic symbol conversion principle, the permutation and combination of this phonetic symbol word string being done group replaces, to produce a plurality of possible phonetic symbol word strings outputs; And

Filter the possible phonetic symbol word string output of this majority;

Above-mentioned filtration comprises the following step at least:

The phonetic symbol of the consecutive identical repetition in this phonetic symbol word string output is filtered, up to surplus next;

If the k in this phonetic symbol word string output or the last phonetic symbol of g phonetic symbol are n, then n is made into η;

If the last phonetic symbol of the ∫ in this phonetic symbol word string output is s, then s ∫ replaces into ∫, if the last phonetic symbol of ∫ is k, then k ∫ replaces into t ∫;

For each this phonetic symbol word string output, seek the individual character that meets; And

Each phonetic symbol word string in this phonetic symbol word string output is derivatized to similar phonetic symbol word string again,, is used to avoid some consonant mute, and the mixing sound is difficult for distinguishing a tin leakage sound that is caused to seek suitable individual character; It is characterized in that:

Above-mentioned deriving at least since second phonetic symbol runs into the situation that continuous consonant is arranged, and once removes this consonant, to form a new phonetic symbol word string;

Also comprise and import at least one pronunciation word string simultaneously;

And also comprise and import at least one wild-character simultaneously.

2. method as claimed in claim 1 is characterized in that, above-mentioned pronunciation word string is and this phonetic symbol word string separate processes.

3. method as claimed in claim 3 is characterized in that, the processing of above-mentioned pronunciation word string comprises the following step at least:

Judge whether this pronunciation word string is consonant entirely, if then can't correct;

Remove non-English and symbol '-' character in addition in this pronunciation word string;

According to English pronunciation principle, set up a pronunciation conversion principle, promptly each phonetic symbol converts corresponding word string or null character (NUL) to according to the phonetic symbol word string at its place;

According to putting in order of this pronunciation word string, and, do the permutation and combination replacement of group according to the pronunciation conversion principle;

The vowel of the consecutive identical repetition of this pronunciation word string is filtered, up to surplus next;

If the k of this pronunciation word string or the last phonetic symbol of g phonetic symbol are n, then n is made into η;

If the last phonetic symbol of the ∫ of this pronunciation word string is s, then s ∫ replaces into ∫, if the last phonetic symbol of ∫ is k, then k ∫ replaces into t ∫;

If " ed " of this pronunciation word string is positioned at last, and if the phonetic symbol in front be p, k, s, ∫, t ∫ θ, f, then the phonetic symbol of ed correspondence changes t into, be false as if aforesaid situation, and last phonetic symbol is not when being not d or t that then the pairing phonetic symbol of ed changes d into;

If the neither establishment of situation of previous step, then the phonetic symbol of ed correspondence changes Id into;

Remove the null character (NUL) in this pronunciation word string;

Go for the phonetic symbol word string of each this pronunciation word string and to seek suitable individual character; And

Each this phonetic symbol word string in this pronunciation word string is derivatized to similar phonetic symbol word string again, removes to seek suitable individual character, its principle of deriving is since second phonetic symbol, runs into the situation that continuous consonant is arranged, and once removes a consonant, forms a new phonetic symbol word string.

4. method as claimed in claim 1 is characterized in that, above-mentioned phonetic symbol word string is and this wild-character separate processes.

5. method as claimed in claim 4 is characterized in that, the processing of above-mentioned wild-character is that this wild-character is replaced into the wildcard phonetic symbol.

6. English spelling inspection method, the input of this inspection method comprises assonance marking-up string, pronunciation word string and wild-character, and this method comprises the following step at least:

According to English pronunciation principle, set up a phonetic symbol conversion principle, be the basis promptly according to a kind of phonetic symbol, all words are converted to possible phonetic symbol word string;

Respectively according to the input sequencing of phonetic symbol word string and this pronunciation word string, and according to this phonetic symbol conversion principle and this pronunciation conversion principle, respectively this phonetic symbol word string and this pronunciation word string are done the permutation and combination replacement of group, to produce most possible phonetic symbol word string outputs and most pronunciation word string outputs, wherein above-mentioned pronunciation word string is and this phonetic symbol word string separate processes; And

Filter these a plurality of possible phonetic symbol word string outputs and the output of this pronunciation word string;

Above-mentioned filtration phonetic symbol word string output comprises the following step at least:

The phonetic symbol of the consecutive identical repetition in this phonetic symbol word string output is filtered, up to surplus next;

If the k in this phonetic symbol word string output or the last phonetic symbol of g phonetic symbol are n, then n is made into η;

If the last phonetic symbol of the ∫ of this pronunciation word string is s, then s ∫ replaces into ∫, if the last phonetic symbol of ∫ is k, then k ∫ replaces into t ∫;

For each this phonetic symbol word string output, seek suitable individual character; And

Each phonetic symbol word string in this phonetic symbol word string output is derivatized to similar phonetic symbol word string again,, is used to avoid some consonant mute, and the mixing sound is difficult for distinguishing a tin leakage sound that is caused to seek suitable individual character; It is characterized in that:

Above-mentioned deriving comprises at least since second phonetic symbol, and running into has continuous consonant person, once removes this consonant, to form a new phonetic symbol word string.

7. method as claimed in claim 6 is characterized in that, more comprises the following step to handle above-mentioned pronunciation word string:

Judge whether this pronunciation word string is consonant entirely, if then can't correct;

Remove non-English and symbol '-' character in addition in this pronunciation word string;

According to English pronunciation principle, set up a pronunciation conversion principle, promptly each phonetic symbol converts corresponding word string or null character (NUL) to according to the phonetic symbol word string at its place;

According to putting in order of this pronunciation word string, and, do the permutation and combination replacement of group according to the pronunciation conversion principle;

The vowel of the consecutive identical repetition of this pronunciation word string is filtered, up to surplus next;

If the k of this pronunciation word string or the last phonetic symbol of g phonetic symbol are n, then n is made into η;

If the last phonetic symbol of the ∫ of this pronunciation word string is s, then s ∫ replaces into ∫, if the last phonetic symbol of ∫ is k, then k ∫ replaces into t ∫;

If " ed " of this pronunciation word string is positioned at last, and if the phonetic symbol in front be p, k, s, ∫, t ∫ θ, f, then the phonetic symbol of ed correspondence changes t into, be false as if aforesaid situation, and last phonetic symbol is not when being not d or t that then the pairing phonetic symbol of ed changes d into;

If the neither establishment of situation of previous step, then the phonetic symbol of ed correspondence changes Id into;

Remove the null character (NUL) in this pronunciation word string;

Go for the phonetic symbol word string of each this pronunciation word string and to seek the individual character that meets; And

Each this phonetic symbol word string in this pronunciation word string is derivatized to similar phonetic symbol word string again, removes to seek suitable individual character, its principle of deriving is since second phonetic symbol, runs into the situation that continuous consonant is arranged, and once removes a consonant, forms a new phonetic symbol word string.

8. method as claimed in claim 6 is characterized in that, above-mentioned phonetic symbol word string is and this wild-character separate processes.

9. method as claimed in claim 8 is characterized in that, the processing of above-mentioned wild-character is that this wild-character is replaced into the wildcard phonetic symbol.

Images