CN1094607C - Intelligent phoneme-shape code input method and application thereof - Google Patents

Abstract

The present invention belongs to the field of computer character input. The character input can be carried out on a standard keyboard by using a plurality of modes, and the character input can also be carried out by using a keyboard additionally with initial consonants and vowels. The present invention is also provided with a telephone number keyboard specially used for the Chinese input, and the telephone number keyboard can be applied to various computer telephone service fields. The present invention is additionally provided with an intelligent input mode which can be widely applied to a plurality of computer character input fields of intelligent input, character proof reading, Chinese character recognition, Chinese speech input, a letter post code of a mailing address instead of a numeral post code, a Chinese display function realization of a BP machine on an English numeral BP machine, etc.

Description

A kind of intelligent phoneme-shape code input method and application thereof

The invention belongs to computword input field.In this field, people have invented many new Chinese character input methods.Some phonetic and configurational code input methods and intelligent input method are wherein arranged.In the input method of the single character of non-whole sentence intelligence input, some method input speed is very fast.For example: this respect has the patented method " simple and easy phonetic-stroke code Chinese character input method " of Mr. Wang Lu and Mr. Xu Huohui invention, and (patent announcement number is 1081772, and statutory status is for authorizing, and its input speed is every word mean code length about 2.0.) and former benefit in the patented method " sound shape stroke integrated encode high-speed Chinese character input method and applied keyboard " or the like of Mr. invention.(patent announcement number is 1039132, and statutory status does not have, and its input speed is that every word mean code length is greatly about about 1.3-1.8.) individual character input mode among the present invention is that a kind of input speed is very fast and convenient and practical, easily learns the easily phonetic and configurational code input method of note.

In states such as Great Britain and Americas, a kind of common English input telephone code keyboard commonly used is arranged.(see Fig. 2 for details..) and be widely used in various compuphone service fields based on the English input technology of the telephone code of this keyboard.But still unmanned proposition of China solves telephone code input in Chinese problem with similar telephone code technology at present.The present invention on the basis that proposes the intelligent phoneme-shape code Chinese character coding input method, designed a kind of suitable input in Chinese the telephone code keyboard (see Fig. 3 for details. and Fig. 4 .) and a series of use telephone code technology technical method of carrying out the telephone code input in Chinese.The present invention also proposes this technology can be widely used in various compuphone service fields, and has specifically noted many application processes wherein.

At present, the technology (see " Journal of Chinese Information Processing ", 1996.2, " a kind of input method based on language understanding--intelligence phonetic letter input method ") of some whole sentence intelligence inputs and the technology that auxiliary literal is proofreaded have been proposed both at home and abroad.And had some products of using whole sentence intelligent input technique and auxiliary literal check and correction technology to come out, for example: with " the logical certainly input " software of the grand light Weir New Tech S. R. L. in Beijing and " input of unexpected rival's intelligence " software and " unexpected rival's check and correction " software of unexpected rival company is a collection of application whole sentence intelligent input technique of representative and the new product of auxiliary literal check and correction technology.But, domestic still unmanned proposition at present is applied directly to the elite of whole sentence intelligent input technique and auxiliary literal check and correction technology and comprises the literal check and correction, Chinese Character Recognition, the Chinese speech input, replace digital postcode with mailing address letter postcode, replace digit phone number with alphabetical telephone number, and the function that on the English digital beeper, realizes the Chinese display beeper or the like many computwords inputs field.The present invention has then proposed many methods of using intelligent phoneme-shape code input method and whole sentence intelligent input technique and auxiliary literal check and correction technology aspect these above-mentioned.

The purpose of this invention is to provide a kind of convenient and practically, easily learn the easily intelligent phoneme-shape code Chinese character input method of note, a kind of telephone code keyboard technique and a kind of adventitious sound vowel keyboard for the input in Chinese use are provided simultaneously.On aforesaid Chinese character coding input method basis, various advanced persons' input in Chinese technology can be applied to extensive fields.Specifically, the telephone code technology can be applied to various compuphone service fields, also intelligent phoneme-shape code input method can be applied to the intelligence input, the literal check and correction, Chinese Character Recognition, the Chinese speech input replaces digital postcode with mailing address letter postcode, replace digit phone number with alphabetical telephone number, and the function that on the English digital beeper, realizes the Chinese display beeper or the like many computwords inputs field.

Technical scheme of the present invention is divided into individual character input mode and the intelligent input mode dual mode of whole sentence.Wherein the individual character input mode is a kind of phonetic-stroke code input mode.Its coding is made up of sound sign indicating number and font code.The principle of regulation is popular easy, easily learns easily note, and encode Chinese characters for computer is shorter.Have five yards types, four yards types, trigram type, inverted order type, shape figure, font code type, multiple input modes such as new the Five-stroke Method and mixed type.Wherein five yards type mode kanji codes are made up of three sound sign indicating numbers and two font codes.The initial that first sound sign indicating number in three sound sign indicating numbers is a Chinese Pin Yin pseudonym, first and last letter that latter two sound sign indicating number is a Chinese phonetic alphabet simple or compound vowel of a Chinese syllable.Wherein, the syllable that belongs to zero initial is handled as simple or compound vowel of a Chinese syllable.(annotate: the someone will be with y[i-] and w[u-] syllable of beginning is included into zero initial.But this input rule is all handled both of these case as initial consonant, and is not included into zero initial.) if simple or compound vowel of a Chinese syllable has only a letter, then the simple or compound vowel of a Chinese syllable of five yards type mode kanji codes part is also only got a letter.Two font codes of five yards type mode kanji codes are after principle is taken into a plurality of parts apart in accordance with regulations with Chinese character, get the Chinese Pin Yin initial of first parts and last component names.The intelligent input mode of whole sentence of intelligent phoneme-shape code input method is on the basis of whole intelligent understanding technology of sentence and auxiliary literal check and correction technology, for the various encoding schemes of individual character input mode or the input multiple encoding scheme of first or preceding several codes or the like of each Chinese character basic announcement sign indicating number wherein only, obtain the Chinese sentence of the appearance possibility maximum that satisfies condition as output with the Markov chain method, provide some possibilities time maximum sentence to be selected simultaneously, and the big part of fallibility provided various compuphone service fields, also intelligent phoneme-shape code input method can be applied to the intelligence input, the literal check and correction, Chinese Character Recognition, the Chinese speech input, replace digital postcode with mailing address letter postcode, replace digit phone number with alphabetical telephone number, and the function that on the English digital beeper, realizes the Chinese display beeper or the like many computwords inputs field.

Technical scheme of the present invention is divided into individual character input mode and the intelligent input mode dual mode of whole sentence.Wherein the individual character input mode is a kind of phonetic-stroke code input mode.Its coding is made up of sound sign indicating number and font code.The principle of regulation is popular easy, easily learns easily note, and encode Chinese characters for computer is shorter.Have five yards types, four yards types, trigram type, inverted order type, shape figure, font code type, multiple input modes such as new the Five-stroke Method and mixed type.Wherein five yards type mode kanji codes are made up of three sound sign indicating numbers and two font codes.The initial that first sound sign indicating number in three sound sign indicating numbers is a Chinese Pin Yin pseudonym, first and last letter that latter two sound sign indicating number is a Chinese phonetic alphabet simple or compound vowel of a Chinese syllable.Wherein, the syllable that belongs to zero initial is handled as simple or compound vowel of a Chinese syllable.(annotate: the someone will be with y[i-] and w[u-] syllable of beginning is included into zero initial.But this input rule is all handled both of these case as initial consonant, and is not included into zero initial.) if simple or compound vowel of a Chinese syllable has only a letter, then the simple or compound vowel of a Chinese syllable of five yards type mode kanji codes part is also only got a letter.Two font codes of five yards type mode kanji codes are after principle is taken into a plurality of parts apart in accordance with regulations with Chinese character, get the Chinese Pin Yin initial of first parts and last component names.The intelligent input mode of whole sentence of intelligent phoneme-shape code input method is on the basis of whole intelligent understanding technology of sentence and auxiliary literal check and correction technology, for the various encoding schemes of individual character input mode or the input multiple encoding scheme of first or preceding several codes or the like of each Chinese character basic announcement sign indicating number wherein only, obtain the Chinese sentence of the appearance possibility maximum that satisfies condition as output with the Markov chain method, provide some possibilities time maximum sentence to be selected simultaneously, and the big part of fallibility is provided mark.The intelligent input mode of whole sentence combines with the integrating words and phrases technology, can increase input, and the correctness of check and correction and identification also can improve the identification computing velocity, increases real-time and Practical Intelligent function.The intelligent input mode of whole sentence can also have and word for word shows Chinese character in real time, automatically memory, self study, revise dictionary adaptively to adapt to user's characteristics, select the error category be used to proofread adaptively, select adaptively and the multiple intelligent function of the type matrix character library used when proofreading and correct the identification Chinese character and specialty characteristics of adaptive user or the like.

The invention allows for a kind of telephone code keyboard technique and corresponding input method of suitable input in Chinese.And point out this keyboard and the corresponding many application of input technology in various compuphone service fields.The intelligent input mode of whole sentence of the present invention can also be applied to the intelligence input, the literal check and correction, Chinese Character Recognition, the Chinese speech input, replace digital postcode with mailing address letter postcode, replace digit phone number with alphabetical telephone number, and the function that on the English digital beeper, realizes the Chinese display beeper or the like many computwords inputs field.

All in all, existing many input methods are merely pursued and are accelerated input speed (promptly reducing every word mean code length) and reduce the repetition rate of coding, but have often ignored convenient and practically, easily learn the easily requirement of note.Therefore though these input methods may surpass on above-mentioned two performance index or methods such as five fonts using always near people in the past and natural code, convenient and practical, easily learn easy note aspect and obviously do not improve.Owing to habitual factor, these new input methods often can't be accepted by people, also can't be applied among the reality like this.The present invention has overcome this shortcoming.Guaranteeing that individual character input mode of the present invention is easily learned easy note aspect and all had clear superiority than aforementioned all methods convenient and practical under the prerequisite that input speed and repetition rate of coding index and five fonts and natural code are more or less the same.Why the natural code method of Mr. Zhou Zhinong invention several years ago holds advantage slightly than five font methods, and main cause has been him with the advantages of sound code method and font code method.The sound sign indicating number of natural code is that the Chinese phonetic alphabet is compressed into the form of being made up of two letters by certain memory rule.The font code of natural code is to be made up of the initial of Chinese character the first, the second parts Chinese title phonetic.Every like this word mean code length is generally less than 4.0.Individual character input mode of the present invention also is a kind of phonetic-stroke code method that the advantages of sound code method and font code method is got up.But the memory rule of its compression Chinese phonetic alphabet is more more convenient and practical than natural code and additive method, easily learns easily note.Therefore on the whole, the present invention has the unexistent advantage of many additive methods.

Compare with the English input technology of the existing telephone code of Great Britain and America, the telephone code input in Chinese technology among the present invention is more suitable for using in the people who uses Chinese, and particularly continent and Singapore can use the people of the Chinese phonetic alphabet.This technology can also combine with the intelligent input mode of whole sentence among the present invention and in the application aspect the Chinese speech input, is widely applied to various compuphone service fields.

Compare with other intelligent input modes, the intelligent input mode input speed of whole sentence of intelligent phoneme-shape code input method is faster under same form.(for example: under full form, the intelligent input mode of whole sentence of the present invention is more a lot of than intelligence phonetic letter method high input speed.Its every word mean code length can reduce by 1-2 code length than intelligence phonetic letter method.) whole sentence intelligent input technique among the present invention combines with the integrating words and phrases technology, can increase input, the correctness of check and correction and identification also can improve computing velocity, increase real-time and Practical Intelligent function.The intelligent input mode of whole sentence among the present invention can also have and word for word shows Chinese character in real time, automatically memory, self study, revise dictionary adaptively to adapt to user's characteristics, select the error category be used to proofread adaptively, select adaptively and the multiple intelligent function of the type matrix character library used when proofreading and correct the identification Chinese character and specialty characteristics of adaptive user or the like.The intelligent input mode of whole sentence among the present invention can also be widely applied to the intelligence input, the literal check and correction, Chinese Character Recognition, the Chinese speech input, replace digital postcode with mailing address letter postcode, replace digit phone number with alphabetical telephone number, and the function that on the English digital beeper, realizes the Chinese display beeper or the like many computwords inputs field.

Below the accompanying drawing among the present invention is made some simple explanations.

Fig. 1. be the adventitious sound vowel keyboard in the intelligent phoneme-shape code input method of the present invention.

Fig. 2. be Great Britain and America's common English input telephone code keyboard commonly used.

Fig. 3. be (I) type input in Chinese telephone code keyboard in the intelligent phoneme-shape code input method of the present invention.

Fig. 4. be (II) type input in Chinese telephone code keyboard in the intelligent phoneme-shape code input method of the present invention.

The individual character input mode of intelligent phoneme-shape code input method is a kind of input method of Chinese character that belongs to phonetic-stroke code, is called for short " phoneme-shape code input method ".Each kanji code of phoneme-shape code input method can be made up of sound sign indicating number and font code.(input method of only using sound sign indicating number or font code is arranged also.) mode of choosing of sound sign indicating number and font code has much characteristic, compares with the various encodes Chinese characters for computer of domestic current, the principle with regulation is popular, and is easy, easy note, advantage such as encode Chinese characters for computer is short.The user only needs first tool Chinese phonetic alphabet knowledge and structure knowledge of Chinese characters just can grasp this input method in very short time.Phoneme-shape code input method has five yards types, four yards types, trigram type, inverted order type, shape figure, four font code types, three font code types, new the Five-stroke Method and mixed type or the like various input.The user can according to circumstances choose any one kind of them in input process.Such as, run into the word of not knowing pronunciation in the secondary character library, just can use four font code types, input methods such as three font code types or new the Five-stroke Method.Five yards type mode particular importances in the phoneme-shape code input method, and the most convenient and practical.Its kanji code is made up of three sound sign indicating numbers and two font codes.The initial that first sound sign indicating number in three sound sign indicating numbers is a Chinese Pin Yin pseudonym, first and last letter that latter two sound sign indicating number is a Chinese phonetic alphabet simple or compound vowel of a Chinese syllable.Wherein, the syllable that belongs to zero initial is handled as simple or compound vowel of a Chinese syllable.(annotate: the someone will be with y[i-] and w[u-] syllable of beginning is included into zero initial.But this input method is all handled both of these case as initial consonant, and is not included into zero initial.) if simple or compound vowel of a Chinese syllable has only a letter, then the simple or compound vowel of a Chinese syllable of five yards type mode kanji codes part is also only got a letter.Two font codes of five yards type modes are after principle is taken into a plurality of parts apart in accordance with regulations with Chinese character, get the Chinese Pin Yin initial of first parts and last component names and form.Following mask body is introduced the basic announcement sign indicating number, basic font code, and the employed kanji code of all kinds method is introduced some extended functions and intelligent function in the phoneme-shape code input method again, again all kinds is done performance evaluation at last.

The basic announcement sign indicating number of phoneme-shape code input method produces on standard Chinese phonetic alphabet basis.Phonetic u in the Chinese phonetic alphabet can represent with v in the various types of input methods of phonetic-stroke code.Initial consonant in Chinese phonetic alphabet part is many is represented to have only ch by single letter, sh, zh are only and are made up of two letters.The initial consonant of phoneme-shape code input method basic announcement sign indicating number partly adopts the initial consonant of being represented by single letter in the Chinese phonetic alphabet, and to ch, sh, and these three initial consonants of being made up of two letters of zh are then only got first letter representation, i.e. ch=c, sh=s, zh=z.The simple or compound vowel of a Chinese syllable part of the Chinese phonetic alphabet generally is made up of 1-4 letter.The simple or compound vowel of a Chinese syllable of phoneme-shape code input method basic announcement sign indicating number partly adopts the simple or compound vowel of a Chinese syllable that is made of 1-2 letter in the Chinese phonetic alphabet, as a, and i, ao, in, ai, an, in etc.; By the simple or compound vowel of a Chinese syllable that 3-4 letter constitutes, then only get first letter and last letter as phoneme-shape code input method basic announcement sign indicating number.Such as: iao=io, uai=ui, ang=ag, uang=ug, uan=un, ian=in or the like.If the phonetic of Chinese character has only simple or compound vowel of a Chinese syllable, there is not initial consonant, the basic announcement sign indicating number of phoneme-shape code input method has only the simple or compound vowel of a Chinese syllable part, as: peace an, like ai.The syllable that promptly belongs to zero initial is handled as simple or compound vowel of a Chinese syllable.(annotate: the someone will be with y[i-] and w[u-] syllable of beginning is included into zero initial.But this input method is all handled both of these case as initial consonant, and is not included into zero initial.) phoneme-shape code input method basic announcement sign indicating number the shortest a letter only arranged, the longest also have only three letters, for example: " two " shuang=sug, " mark " piao=pio, " former " yuan=yun, " shop " dian=din.(annotate: with c, s, z and ch, sh, these two groups of initial consonants of zh merge, and use c, s, one group of letter representation of z.The people that this one side also is some localism area for convenience can not distinguish this two groups of initial consonants.Also have font code to distinguish in sound sign indicating number back when importing each Chinese character, therefore generally can not cause the repeated code word.)

Hanzi font library in the computing machine is divided into one-level character library and secondary character library.The one-level character library is Chinese characters in common use, and the secondary character library is the Chinese character that is of little use, and generally mostly is the more and complicated Chinese character of stroke.According to this characteristics, the basic font code of phoneme-shape code input method is made up of two font codes, and it is that structure according to Chinese character is provided with.The structure of Chinese character can be analyzed and be two kinds of single character and combinde rqdical characters.Single character is a whole word, can not analyze out.Combinde rqdical character is made up of plural composition.These compositions are single character a bit, and some is radical or the stroke of independently not using as word.No matter be single character or combinde rqdical character, phoneme-shape code input method is torn Chinese character open and is a plurality of parts.Press Chinese character order of writing strokes (promptly from left to right, from top to bottom, from outside to inside, from the centre to both sides.) situation of looking the first stroke and finishing touch determines two basic font codes.

The multiple principle of having divined by means of characters is as follows:

One. only word principle: when the first stroke or finishing touch during with relevant stroke formation single character, first letter of just getting this single character Chinese phonetic alphabet is as font code.As: " compassion " the first stroke and relevant stroke formation single character " non-", its font code is f, and finishing touch and relevant stroke constitute single character " heart ", and its font code is x, and like this, the basic font code of " compassion " is exactly fx.

Two. the radical principle: the first stroke or finishing touch and relevant stroke only constitute radical, first letter (for details, see the appendix two) of just getting the radical title Chinese phonetic alphabet is font code, as: the first stroke in " river " constitutes radical " Rui " with relevant stroke, its Chinese title is " 3 water ", and first the alphabetical s of the Chinese phonetic alphabet that gets " three " makes font code; Finishing touch constitutes single character " worker " with relevant stroke, makes code with g.The font code in " river " is sg.And for example: " beating " the first stroke constitutes radical " Rolling " with relevant stroke, is " by the handle ", and first the alphabetical t of the Chinese phonetic alphabet that gets " carrying " makes font code; Finishing touch constitutes single character " fourth " with relevant stroke, makes code with d.The font code of " beating " is td.

Three. the stroke principle: the first stroke or finishing touch promptly do not constitute single character with relevant stroke, do not constitute radical yet, and font code (for details, see the appendix two) made in title (point, horizontal, vertical, left-falling stroke, right-falling stroke, folding etc.) first letter of the Chinese phonetic alphabet of just getting this stroke.As " little " the first stroke is perpendicular colluding, and finishing touch is a little, just uses sd; " sorrow " the first stroke is a little, and finishing touch is to press down, and just uses dn; " river " is with erecting ss.

Four. get big principle: the divine by means of characters principle of getting parts of phoneme-shape code input method is got greatly, does not get little; Get popularly, common, do not get ancient character and rare.Such as: the first stroke of " gloomy " constitutes single character " wood " with relevant stroke, makes code with m.Finishing touch both constituted " wood ", but also constituted " woods " (seeing one).Press big principle, back one font code is only got " woods ", does not get " wood ", and then the font code of " gloomy " is just used ml.

Five. unisonance is evaded principle: first is alphabetical when identical if the sound sign indicating number of the font code of parts and this word is arranged, and the parts of this font code representative are just again toward tearing open to avoid repeated code for a short time.As " removing ", press big principle and should be " Rolling+as ", still " as " identical with the sound of " removing ", therefore " to remove " just to tear open and be " Rolling+again ", the font code of " removing " should be ty.

Six. similar shape is evaded principle: when second font code that is taken out as stated above is identical with first font code, if can split into three with upper-part, then get two font codes of first parts and second parts.If can only tear two parts open, then font code is constant.Can split into " king+Dian+Pie+king " as " class ", by the first five principle, basic font code should be king king's two font codes, i.e. ww.But by principle six, should get preceding two parts " king+Dian ", its font code is wd.And for example " woods " can only split into two parts " wood+wood ", and its font code is mm.

The encode Chinese characters for computer of five yards type input methods generally by the longest be that three sound sign indicating numbers add that two font codes form.Wherein three sound sign indicating numbers are foregoing basic announcement sign indicating number, and two font codes are foregoing basic font code.

Hanzi font library in the computing machine is divided into one-level character library and secondary character library.Chinese character in the one-level character library is everyday character.Chinese character in the secondary character library is the Chinese character that is of little use.Usually, mostly to be stroke many and complexity and common people are difficult to determine the Chinese character of its pronunciation for the secondary character library.According to these characteristics, five yards type input methods have added another four font code type input method to the Chinese character in the secondary character library, and both optional one can be imported the Chinese character in the secondary character library.

Four font code type input methods are only used four font codes, do not use the sound sign indicating number, do not use so that know the people of Chinese character pronunciation.The method of divining by means of characters of four font code type input methods is as follows:

It is four parts that every word is all torn open.The first step is determined earlier two parts and code thereof by the basic font code principle of divining by means of characters fully, is that the parts more than the stroke split into two parts by aforementioned principle more then with big parts, determines its code.As " gal ", the first step is: " Ren "+" adding " (dj), second step tear open " adding " be " power "+" mouth " (lk), the font code of " gal " is djlk.If two component sizes of tearing open for the first time equate that second step was just torn back one parts open.

The encode Chinese characters for computer of four yards type input methods is generally the longest to be that two sound sign indicating numbers add two font codes.If the basic announcement sign indicating number is less than two, the sound sign indicating number of four yards type input methods is identical with the basic announcement sign indicating number, and more than two, the sound sign indicating number of four yards type input methods is preceding two sound sign indicating numbers of basic announcement sign indicating number as if the basic announcement sign indicating number, and two font codes wherein then are basic font code.Foregoing four font code type input methods are equally applicable to this section.

The encode Chinese characters for computer of trigram type input method is generally a sound sign indicating number and adds two font codes.The sound sign indicating number is first sound sign indicating number of basic announcement sign indicating number, and font code is basic font code.Foregoing four font code type methods are removed the three font code type input methods that three font codes behind the 3rd font code are formed.

Inverted order type input method adds that by two font codes three sound sign indicating numbers form.Mainly be applicable among the UCDOS and WINDOWS95 of prompt facility gradually.Be characterized in: the basic font code of two font codes compositions of input earlier, import the basic announcement sign indicating number that three sound sign indicating numbers are formed again, promptly with five yards type reversed in order.Compare with five yards types, this method input speed is faster, but it is convenient like that naturally to be not so good as five yards types.Four font code type input methods in § 2.4 joints are equally applicable to this section.

Shape figure input method generally adds that by a sound sign indicating number three font codes form.The sound sign indicating number is first letter in the basic announcement sign indicating number.The method of tearing open of three font codes is as follows: the first step is fully by the principle of divining by means of characters of basic font code, determine two parts and code thereof earlier, then the many parts of stroke are wherein splitted into two parts by its principle again, only get its last parts and determine its code, such three codes are formed three font codes.(if two component sizes of being torn open for the first time equate, just tear back parts open).For example: " example " is removable to be " Ren " and " row ", and then " row " torn open is " bad " and " Dao ", only gets " Dao ", and therefore, the code of " example " is dll.Equally, " Guo " is removable for " enjoying " and " Fu ", and taking " enjoying " again apart is " Dian " and " son " and only get " son ", and its code is xez.(annotate: the font code part of shape figure input method can be similar with foregoing three font code type input methods.)

Shape figure input method is used new five-stroke character input method for the Chinese character in the secondary character library, it is torn method open and gets its first three parts and last parts for splitting out a plurality of parts by the Five-stroke Method method of tearing open earlier, and the method by the basic font code of generation noted earlier generates new five code word type input method codes again.As " seize by force " (sound ji), the code of the Five-stroke Method method is dskj, and each parts that its correspondence splits out are " big fourth mouth Dao ", and then the code of new five-stroke character input method correspondence is ddkl.

The various mixed methods that also have several or whole mixing uses in will above-mentioned five types in the phoneme-shape code input method.In these mixed methods, can choose certain several method that is blended in wherein wantonly and carry out the Chinese character input, can export correct Chinese character.The user who goes for various different needs like this uses.Because various mixed methods are comparatively complicated, the various files of formation are too assorted too much, temporary transient now type of recommending use.

In phoneme-shape code input method, design has multiple fault tolerance.Remove WINDOWS95, outside the fault tolerances such as universal key among WINDOWS3.1 and the UCDOS or fault-tolerant key, phoneme-shape code input method also has the odd encoder fault tolerance of oneself.When a Chinese character has multiple differently when tearing method open naturally, can choose any one kind of them and tear the method input open.In addition, phoneme-shape code input method also has the special fault-tolerant key function of oneself.Because alphabetical i, u, v can not appear at the initial place of the Chinese phonetic alphabet, therefore necessarily can not be as the first sound sign indicating number and each font code of phoneme-shape code input method encode Chinese characters for computer.When you can't determine the first sound sign indicating number of a Chinese character and when divining by means of characters font code, can use by alphabetical i, u, the fault-tolerant key that v forms (also claim fuzzy key or in and key) replace this code.Special fault-tolerant key has three.First is a letter " i ", represents the font code that those can't determine type.Second and the 3rd is letter " u " and " v ", represents the font code of two class fixed types respectively.

In addition, also have hundreds of brevity code words to import in the phoneme-shape code input method with simplified way.The input method of each brevity code word is: as long as these two character codes of first font code of first sound sign indicating number in this Chinese character basic announcement sign indicating number of input and basic font code.

One. fault tolerance: when certain Chinese character have two or more different when tearing method open naturally, the method for tearing open of can choosing any one kind of them.

Two. spelling function: also can be during input sound sign indicating number by spelling method input sound sign indicating number.

Three. the word-building function: input function simplified in phrase.

Four. increase the neologisms library facility: (slightly.)

Five. select installation function: (slightly.)

Six. phoneme-shape code input method has WINDOWS95, most of advanced function that Chinese character coding input method had under WINDOWS3.1 and the UCDOS.

One. the basic word-building rule of intelligent phoneme-shape code input method individual character input mode is:

(1) two-character word: press the P11+P21+P12+P22 mosaic.Piecing together as " hello " is nhia.

(2) three words: triliteral first initial addition.Piecing together as " Communist Party " is gcd.

The above speech of (three) four words: first three prefix letter adds last prefix letter.Piecing together as " at a tremendous pace " is yrqd, and " Marxism-Leninism " pieces together is mksy.

Two. the word-building rule of foregoing five yards type input methods is for expanding word-building rule:

(1) two-character word: press P11+P12+P21+P22 (can choose one wantonly) with basic word-building two-character word method.Piecing together as " hello " is niha.

The together basic word-building of speech that (two) three words are above.

Three. foregoing trigram type input method word-building rule is with among (one) in the basic word-building, (two), (three)

Three codes remove and get final product.Other input method all adopts basic word-building rule.

Five yards type input modes in the phoneme-shape code input method claim convenience type again, are the most convenient and practical phoneme-shape code input methods, and it is the most simple and convenient and easy to study, are easy to promote very much, and are practical.But code length is 2,3,4,5, and maximum code length is 5, and code length is than trigram type (code length is 3), and four yards types (code length is 4) all omit long.It is more lower slightly than inverted order type efficient, but more natural convenience.Its character library repetition rate of coding is about 8%, and it is higher than shape figure (the character library repetition rate of coding can be low to moderate be about 1%), but will hang down than trigram type (the character library repetition rate of coding is up to 20%--50%), is moderate performance, is suitable for particularly a kind of method of beginner of users.

Trigram type code length is the shortest, in fact almost reaches capacity, and can't provide the input method of follow-on more short code.And its repetition rate of coding is very not high, and＞=50% word can be selected.In addition, other word, repeated code word＜=10% in the most repeated code words＜=10, particularly one-level GB character library, one-level secondary are added up the overwhelming majority also＜=10%.Like this, most everyday characters are directly done once to select to get final product without page turning, are particularly suitable for computer utility, can import Chinese character alternately fast with screen, general maximum four lower keyboards of making a call to of every word, and a large amount of situation needs only＜=3 lower keyboards.Stroke minimum and the input the fastest.

Four yards type code lengths are generally the longest to be 4, and identical with the Five-stroke Method maximum code length, comparatively simple and convenient and easy to study, the sound sign indicating number is preceding two sign indicating numbers of five yards type sound sign indicating numbers, and font code is the same with five yards types, and the repetition rate of coding increases few than five yards types, and comfort level is also similar.But maximum code length is shorter, and performance is slightly improved.Be comfort level, a kind of method of important indicators such as the maximum code length and the repetition rate of coding between five yards types and trigram, moderate performance is fit to promote, and can stand severe tests in market.

The inverted order type is that the sound sign indicating number of five yards types and font code input sequence are turned around, and two basic font codes of input are imported three basic announcement sign indicating numbers more earlier.Because five yards types are input sound sign indicating numbers earlier, when pointing out gradually,, can avoid with the font code comparatively situation of difficult of divining by means of characters as can not importing font code with options button by the sound sign indicating number.If import font code earlier with the inverted order type, must divine by means of characters with font code, so comparatively difficulty and not nature, this respect comfort level is not as five yards types.But, so after often importing preceding four codings of inverted order type, just had only a unique individual character with the code word storehouse because font code is cut apart latter two sound code efficiency height of efficiency ratio.At this moment can strike space bar or enter key and import this word.Like this, defeated word mean code length shortens greatly, and smaller or equal to four, five code lengths, so the inverted order method also has certain distinct advantages.

Shape figure input mode code length in the phoneme-shape code input method is 2,3,4, and general maximum code length is 4, and code length is placed in the middle, between trigram type (code length is 3) and five yards types (code length is 5).Its moderate performance, code length is the same with famous the Five-stroke Method method that certain customers habitually practise.And the repetition rate of coding is very low to be about 1%, and easily learns than the Five-stroke Method is convenient, only needs several simple rules of association to get final product, and is a kind of input method of function admirable.But it will import three font codes, divine by means of characters than the trigram type, and four yards types, five yards types and inverted order type etc. are more difficult and the time is longer, are a kind of phoneme-shape code input methods that more biases toward font code.Therefore, it is than trigram type, four yards types, and five yards types and inverted order type etc. more find it difficult to learn difficult the popularization.

At first, for the vowel ü in the Chinese phonetic alphabet, when using in its independent use or with other monograms,, both can represent also can represent with u with v.Also can design and to represent maybe the scheme that can only represent with v with u.Like this, the representation scheme of Chinese phonetic alphabet vowel ü just can have many kinds.Wherein best scheme is both can represent also the scheme that can represent with v with u.

In addition, because alphabetical i, u, v can not appear at the initial place of the Chinese phonetic alphabet, therefore necessarily can not be as the first sound sign indicating number and each font code of encode Chinese characters for computer.Like this, can use letter key i, u, v is as the fault-tolerant especially key of the first sound sign indicating number and each font code.(see § 2.11 for details.) wherein i can represent to determine the code of type, u can represent the code of the first kind, v can represent the code of second type.

Adventitious sound vowel keyboard such as Fig. 1..Each Chinese phonetic alphabet single-letter initial consonant and zero initial on QWERTY keyboard, also have ch, sh, zh totally three initial consonants can not be with the letter key direct representations on the QWERTY keyboard.In the simple or compound vowel of a Chinese syllable and zero initial of the Chinese phonetic alphabet, remove a that can represent on the QWERTY keyboard, e, i, o, u is outside the ü, (wherein ü can represent with letter key u or v) also has ai, an, ang, ao, ei, en, eng, er, ia, ian, iang, iao, ie, in, ing, iong, iu, ong, ou, ua, uai, uan, ü an, uang, ü e, ui, un, ü n, many simple or compound vowel of a Chinese syllable of uo or the like or zero initial can not be with the letter key direct representations on the QWERTY keyboard.(annotate: the someone will be with y[i-] and w[u-] syllable of beginning is included into zero initial, this input rule all as the initial consonant processing, and is not included into zero initial with both of these case.)

For all these initial consonants that can not use the QWERTY keyboard direct representation, simple or compound vowel of a Chinese syllable and zero initial are added the vowel ü in the Chinese phonetic alphabet, produce the adventitious sound vowel keyboard of being made up of corresponding sound final key.(see Fig. 1 for details..) at this moment, as long as after QWERTY keyboard and adventitious sound vowel keyboard all connected with main frame, button was directly imported any initial consonant in the Chinese phonetic alphabet in the above, simple or compound vowel of a Chinese syllable and zero initial are as long as the Chinese phonetic alphabet of each Chinese character strikes two lower keyboards at most.This keyboard uses simple and convenient, and is directly perceived practical, easily learn easily note, and input speed is very fast.Therefore this is the very strong computing machine input in Chinese equipment of a kind of practicality.

In states such as Great Britain and Americas, a kind of common English input telephone code keyboard commonly used is arranged.(see Fig. 2 for details..) and be widely used in various compuphone service fields based on the English input technology of the telephone code of this keyboard.But still unmanned proposition of China solves telephone code input in Chinese problem with similar telephone code technology at present.The present invention has designed the telephone code keyboard of a suitable input in Chinese and the method that a series of use telephone code technology is carried out the telephone code input in Chinese on the basis that proposes the intelligent phoneme-shape code Chinese character coding input method.The present invention also proposes this technology can be widely used in various compuphone service fields, and has specifically noted many application processes wherein.

Compare with the English input technology of the existing telephone code of Great Britain and America, the telephone code input in Chinese technology among the present invention is more suitable for using in the people who uses Chinese, and particularly continent and Singapore can use the people of the Chinese phonetic alphabet.This technology can also reach with the intelligent input mode of whole sentence among the present invention and combine in the application aspect the phonetic entry, is widely applied to various compuphone service fields.

As Fig. 3. and Fig. 4. shown in, on 12 common key telephone keypads, on nine numerical keys of 1-9, every key stamps zero successively, three or four corresponding letters or symbol.To stamp 26 English alphabets and some symbols altogether, for example: the vowel sign in the Chinese phonetic alphabet " ü " and symbol ". " or the like.Wherein, can replace alphabetical v, perhaps replace symbol ". " with ü with the vowel ü in the Chinese phonetic alphabet.This input in Chinese telephone code keyboard can have a variety of forms, Fig. 3. and Fig. 4. listed for two kinds of comparatively practical forms of input in Chinese.Fig. 3. be (I) type input in Chinese telephone code keyboard, Fig. 4. be (II) type input in Chinese telephone code keyboard.

During input Chinese, press earlier the sound sign indicating number of Chinese phonetic alphabet decision Chinese character, principle determines font code in accordance with regulations again, imports by these codes then.Also can only import first code or preceding several code in these codes.If the input method among use the present invention can significantly reduce code length, and reduce the repetition rate of coding.Use also very simple and conveniently, directly perceived practical, easily learn easily note.When carrying out input in Chinese with the telephone code technology, the method according to special regulation determines the alphanumeric codes that 26 letters are formed earlier, presses the corresponding phone numbers code of corresponding numerical key input again, at last by the corresponding Chinese character option of these codelookups.Run into when repeated code is arranged, can utilize and play relevant recording, relevant Chinese character title, phrase or Chinese option are selected or imported to the interactive mode of asking the user to select.Carry out mutual total amount for reducing playback, use the less input method of the repetition rate of coding.Because the Chinese character common phrase generally has only several ten thousand, often only use part Chinese phrase wherein when carrying out input in Chinese again with the telephone code technology.So can determine the phrase alphanumeric codes that several codes connect to form before each Chinese character according to the word-building rule described in the § 2.12, press corresponding numerical key again.When the item to be checked of database only uses part Chinese character phrase, for example use when being less than 30000 phrases, use four yards above-mentioned phrase codes, total about 10000 kinds of possibilities, the repeated code number generally is less than 10.Like this, carrying out the mutual total amount of selecting of playback is acceptable.If what use is toy data base (application of many practicalities belongs to this type of), employed Chinese character phrase only is equivalent to several thousand magnitude or still less, use said method (i.e. four yards phrase codes) then can accomplish almost not have repeated code, also just selected alternately with playback hardly.In corresponding software engineering, can set up a database.Wherein each data has four yards phrase code items of telephone code that to be checked of Chinese phrase and combines by nine numerals of 1-9 and the item of information that some are corresponding.After each use telephone sound card and corresponding software receive the telephone code input, can use the data base querying technology, find four yards phrase code items of telephone code of all input code appointments, after carrying out the repeated code selection through playback again, can find out data item and the corresponding information thereof that to look for definitely, just can play these information to the user at last by phone.So just can utilize above-mentioned input in Chinese telephone code technology to carry out input in Chinese, the multiple application of data base querying and mutual selection or the like.So above-mentioned input in Chinese telephone code technology can be widely applied to various compuphone service fields.

At first, input in Chinese telephone code technology recited above can be applied to the automatic message service of beeper aspect.127 beeper automatic paging business have been opened in the more existing cities of China now.But 127 can only automatic paging, can not leave a message automatically.Use input in Chinese telephone code technology recited above, can be after input automatic message-leaving function number and assistant-searching catchword, again with the brief message information of telephone code code input of correspondence.(for example: the surname of called person, the name of called person, brief message, time, place or the like.Coding can be referring to rearmost encoding scheme in detail.) the automatic information of discerning input of beeper paging centers meeting, send corresponding signal and signal is shown on user's beeper.The numeral beeper, digital English beeper and Chinese display beeper all can use this technology.This The Application of Technology can improve the robotization service level of beeper industry greatly.

The second, input in Chinese telephone code technology recited above can be applied to 114 directory enquiry unmanned automated management systems.Can use this input in Chinese telephone code technology to import on phone by the name of enquiring telephone number (or organization), after mutual the selection, system can find out by the directory enquiry sign indicating number automatically and play corresponding recording and quote by the directory enquiry sign indicating number to the user.So just can save 114 directory enquiry personnel, save manpower, save expense, improve automatization level.Further, can use above-mentioned telephone code technology and realize unattended switch system with name or organization automatic telephone switching.Can in a large amount of band branch exchange machine system of China, at first use this technology, set up a band branch exchange machine system that uses above-mentioned input in Chinese telephone code technology with name or organization automatic telephone switching, replace operator, connect phone automatically according to calling name or organization that the caller imported.For example: be the phone in logical Shoudu Iron and Steel Co two workshops, can put through the Shoudu Iron and Steel Co exchange earlier, with the telephone code code of above-mentioned telephone code technology input " two workshops " three words, system just can be switched to phone Shoudu Iron and Steel Co two workshops automatically then.

Similarly replace the application of digital postcode with address letter postcode, can be on above-mentioned technical foundation, set up and a kind ofly replace digit phone number with alphabetical telephone number, use the switch system of name (or organization) automatic telephone switching.Can realize that like this user moves or need not notify other people when changing telephone number, and other people still can put through the advanced function of this user's phone.Can also realize that user's telephone number is secret, connect in limited time, change the hotline number, stay the multiple function of hotline number and voice mail or the like when going on business.Also can produce and a kind ofly use above-mentioned input in Chinese telephone code technology or directly use the letter key of Chinese pin yin dish to carry out input in Chinese, direct input alphabet telephone number is transferred to the phone of the digit phone number of alphabetical telephone number correspondence then automatically by the electronic installation in the phone.Can a small display screen be installed on this phone, check display screen, revise at any time and error recovery while the user can import.After after confirming, just formally transfer to corresponding digit phone number.Wherein, the alphabetical telephone number of digit phone number correspondence can be by user oneself input decision.This phone must be practical, can have certain market.Can use computer technology such as call voice card technique on the phone that links together with computing machine, to realize this function, also can produce a kind of comparatively cheap telephone for special use specially.

Can at first use a kind of alphabetical telephone code trunk code telephone number that utilizes input in Chinese telephone code keyboard technique.This novel alphabetical telephone code trunk code telephone number can be divided into two kinds.First kind is novel short code letter telephone code trunk code telephone number, it is made up of three numerals, preceding two numerals are input in Chinese telephone code numerals that this toll telephone office is wanted the initial correspondence of preceding two the word Chinese phonetic alphabet of regional Chinese title, and third digit is the repeated code sequence number.This short code letter telephone code trunk code telephone number can be applied to tens or a hundreds of main cities of China.For example: according to Fig. 3. shown in (I) type input in Chinese telephone code keyboard on the telephone code rule stipulated, the short code of Beijing letter telephone code trunk code telephone number can represent with 251 or 250, and the short code letter telephone code trunk code telephone number of Hohhot City ,Inner Mongolia Autonomous Region then can add that the 3rd repeated code rank-numeral represent with 33.Use like this and remember all very simple and convenient.Second kind is novel long code word base telephone code length way area code telephone number, generally is applied to some areas, zonule in each big zone.Can not use the area of first kind of number generally can both use second kind of number.It is by five or six digital compositions, preceding two numerals are provinces that this toll telephone office is wanted the place, zonule, the city, the input in Chinese telephone code numeral of the initial correspondence of preceding two the word Chinese phonetic alphabet of the big regional Chinese title of autonomous region or the like, the 3rd and fourth digit are the input in Chinese telephone code numerals that this toll telephone office is wanted the initial correspondence of preceding two the word Chinese phonetic alphabet of zonule Chinese title, and the 5th or the 5th and the 6th numeral are the repeated code sequence numbers.Repeated code is less and when being less than 10, the numbers that can use five numerals to form.Repeated code is more and more than 10 but when being less than 100, the numbers that can use six numerals to form.General repeated code can be more than 10, more can be more than 100.Usually, can arrange the repeated code sequence number according to the sum of having a telephone installed in each area, many sequence numbers of having a telephone installed come the front.For example: the long code word base telephone code length of Inner Mongolia Autonomous Region Erenhot City way area code telephone number can add that the repeated code sequence number represents with 6635.This like this toll number is fully can be received no longer than six bit digital.As long as everywhere phone generally uses Fig. 3. shown in (I) type input in Chinese telephone code keyboard and corresponding telephone sign indicating number technology, this alphanumeric toll number uses and remembers can be very simple and convenient.It can save the trouble of many memories.

Suggestion stipulates that artificially the character code of toll number area code correspondence is identical with this area's letter postcode.Just can accomplish this point as long as the regulation principle of the two is consistent.Using like this can be more convenient and practical.

The 3rd, above-mentioned telephone code technology can use help to set up a kind of valuable and easily lose the LR coded system of article and the phone on this coding and above-mentioned telephone code technical foundation, the set up inquiry system of reporting lost property to the authorities.Can set up one or more Register, any valuable and article that easily lose heart are in these applied for registration of one for losing the LR number that report is used.Register will set up corresponding data archival storehouse, all persons of applying for registration of are distributed the LR number successively, note the LR number of all approvals, corresponding applicant, article owners etc. for information about, and guaranteeing that each article only uses a LR number, any two different article all can not double sign.Register or article owner will entrust specialized agency to stamp on article then, write or engrave corresponding LR number.Like this when this article lost, the insider just can report lost property to the authorities by the phone that phone is set up to the various places inquiry system of reporting lost property to the authorities, and report is lost the LR number of article and other for information about.In phone is reported lost property to the authorities inquiry system, can use above-mentioned telephone code technology and set up unattended automated management system, carry out input in Chinese automatically, numeral input and mutual the selection.Set up a data file store simultaneously, note all current LR numbers that are in the state of reporting lost property to the authorities.All in the storehouse, increase a record at every turn when the someone reports lost property to the authorities, write down corresponding LR number.Article find or cancel report lost property to the authorities after, can cancel corresponding record.Other people purchase article or suspect by other approach certain article may incoming road under timing, if these article are printed on, with or be carved with the LR number, the phone of just can the making a telephone call to inquiry system of reporting lost property to the authorities is inquired about these number article and whether is the state of reporting lost property to the authorities.If report lost property to the authorities state, can take corresponding measure, avoid oneself being subjected to any loss, also can help the owner who loses article to find relevant clue of losing article, and find the criminal as early as possible.If not the state of reporting lost property to the authorities, then can relieved purchase or no longer further investigation.Can protect popular interests like this, particularly lose owner with the businessman's of purchase article of article interests.This can make the popular bigger sense of security that produces, and stabilizes society, and suppresses crime, serve the general public, and be a good job of benefiting the nation and the people.

The 4th, for various database inquiry systems, all can use above-mentioned telephone code technology to realize being undertaken the function of data base querying by phone.Combine with Internet technology, can realize a kind of suitable China's actual conditions, convenient and cheap, the Internet system that can conduct interviews by phone.

At last, can use above-mentioned telephone code technology to carry out the far-end control and the far-end operation of computing machine with phone.And it is convenient and practical to operate, and friendly interface is fit to Chinese and uses.Further, above-mentioned telephone code technology and Internet technology combine, and can realize inquiring about for information about function by phone on the internet.The further revolution of the telephony of videophone innovation can be linked the easily a kind of of diverse network, cheap and practical terminal device with phone is become.

Introduce a kind of fast Chinese character input method of understanding based on " statement " below.This method has been utilized contextual correlativity, realizes the automatic conversion that briefly is encoded to Chinese character of Chinese character.The user only need import corresponding Chinese character and briefly encode, do not select Chinese character by hand, based on context system just provides corresponding Chinese character automatically, and according to the input content change result is dynamically adjusted in the scope of whole statement, guarantees the correct of statement at any time.This method has reduced stroke, and can realize touch system basically, has greatly improved the speed of Chinese character input, and just can grasp with learning training hardly.(see list of references [7.] for details.)

It is a long-standing problem that Chinese character is input in the computing machine, although proposed nearly thousand kinds of Chinese character input methods (encoding scheme) at present both at home and abroad, does not accomplish also that all can import Chinese character fast can grasp again at an easy rate.Existing method can roughly be divided into two classes: a class is around strokes of Chinese characters encoding, the another kind of coding that is various based on pronunciation (phonetic).Two class methods respectively have characteristics, and last class methods typically have five graphemic codes etc. based on " divining by means of characters ".The characteristics of this method are that repeated code is few, and high input speed can be realized touch system, therefore are fit to professional typing personnel and use.But the thing that will remember because of it a lot (with five be example, 227 characters and a lot of input rules are arranged), need the training of flower long time could use, make common people be difficult to grasp.Even learned, if a period of time need not, just be easy to not familiarly, add when using these input methods, consider how to divine by means of characters, do not meet the custom that people use language, in fact can't accomplish the input of completing.Back one class methods are as long as can just can grasp by phonetic, and the people of meeting phonetic is a lot.More since voice to be that the mankind transmit information each other the most natural, most convenient and the most effective form, so meet the custom that people use language, the input of accomplishing easily to complete based on the input method of pronunciation.So,, still have many people using it although this class methods speed is very slow at present.But the input speed just because of it is too slow, and it is not suitable for importing long article.

Someone thinks that existing phonetics input method speed is slow, mainly is that actually this is not so because stroke is many.Statistics to a large amount of language materials shows, have only 3.06 letters with average each sound of the spelling of national Scheme for the Chinese Phonetic Alphabet coding, after considering that when repeated code sorts word frequently, import an average stroke of Chinese character below five times, if and use the current average alphabet length of each pronunciation of simplicity method is 2.11 (being as the criterion with WPS), import a Chinese character stroke below 3.5 times, even this speed does not reach the speed of methods such as five, also suitable basically, can be too not slow.And in fact, speed that present various phonetics input methods can provide and last class methods are not on same magnitude.Cause this result's reason to be that existing phonetics input method is behind the pronunciation of a word of input, also must from numerous phonetically similar words, manually select desired word, at this moment the user will sweep whole presenting bank, even also wants page turning, and this has just influenced input speed greatly.And when notice is placed on presenting bank, can't realize touch system.Therefore, the efficient key that improves phonetics input method is to become manual word selection and is automatic word selection.

Chinese has the characteristic of a sound multiword, and in other words repeated code is a lot.The Chinese pronunciation of including with Xinhua dictionary is as the criterion, and has 412 (not considering tone), and secondary GB Chinese character is equivalent to have 7536 words after considering a word multitone, corresponding 18.29 the GB Chinese characters of average every kind of pronunciation.Therefore, solve sound multiword easy thing by no means.

Phonetics input method based on speech is also arranged at present, by the speech typing or provide prompting, go a step further, on Cheng Du necessarily, solved the problem of word selection than in the past as employing such as association's coding and Two bors d's oeuveres double-tone input methods, but still very far away from the target of touch system, mainly show:

1. what will run in a large number in the Chinese language text of reality is a words, as: "Yes", " with ", " " or the like, these still need the manual Chinese character of selecting.

2. can't handle basically the above speech of three words.

3., need manual the selection even also there is the homonym problem in two words that can better handle.

4. in fact the user does not know which speech is included, can directly press the speech input, and which then cannot.Therefore, after regular meeting occurs press the speech input Pinyin, find in the dictionary not this speech, pressed many keys for no reason more.

Although two class methods are all in the convenience of updating with raising speed and use, from the angle of coding research, they have a common weak point, have ignored the influence of natural language context dependence to coding exactly.The research method of whole sentence intelligent input method is different with coding research method in the past.They are that information theory is used for encode Chinese characters for computer, in short doing as a whole research, utilize the context dependence of sentence to reduce code length, be to realize that by sentence comprehension pinyin string or the brief coded strings of Chinese character arrive the automatic conversion of Chinese character string specifically, thereby save the process of manual word selection.The whole sentence intelligent input method that this section proposes has overcome the weak point of aforementioned input method, be a kind of can touch system, meet people's speech habits, easily the rapid input method of grasping.The user only need import the brief coding of Chinese character and not need manual word selection, just can import article, input speed can near in addition surpass quick input method such as five fonts.

From information-theoretical angle, the process of input Chinese character is actually the people provides process from information to computing machine.As long as enough information is provided, just can uniquely determine a Chinese character.Therefore the research coding at first will be understood and determine how many quantity of information of the minimum needs of Chinese character in one piece of article, and in information theory, it is equal to the quantity of information that each Chinese character on average comprises.

If the Chinese character in one piece of article is regarded as independent equiprobable, because secondary GB Chinese character has 6724, need on average to represent a Chinese character that we call an average information that Chinese character comprised to it with 12.7bit.If with 26 alphabetic codings, each alphabetical information amount is 4.7bit, therefore on average the shortlyest be encoded to 2.7 letters.And if consider frequency difference and the contextual correlativity of natural language that different Chinese character in fact occurs in article, distinguish the required bit number of each Chinese character and will descend a lot.After the difference of having considered on each Chinese character frequency, distinguish the average 9.6bit (seeing [1.]) that only needs of a Chinese character.After having considered context dependence, this value also will descend, although still do not know what are reduced to for Chinese at present, but from the situation of English research and the situation of relevant speech understanding, bit number can reduce by 1/3 at least again and (see [2,3,4]), promptly about 6 bits.The shortest coding is on average about 1.3 letters.

What provide above is theoretical value, and when actual coding, owing to the convenience that will consider to use, average code length is worth greater than this.The purpose of research encode Chinese characters for computer is to provide a kind of easy to use, the average few coding of stroke.From top analysis as can be seen, reduce length and can set about from two aspects, the first, study the structure or the pronunciation of each word, reduce code length to each word; The second, utilize contextual correlativity, reduce the redundance of coding.What research was in the past almost walked all is article one, and individual other considered the group speech.At present various methods based on " divining by means of characters " are near the limit of 2.7 keys, can only reduce zero point several times or even write an article on the keystroke several times zero point zero each other.If unfavorablely use contextual correlativity, frequently, at most also can only accomplish about 2 keys, one word even considered word.And the coding that obtains like this is to be difficult to memory.

Here the input method of Ti Chuing is second approach above adopting, and promptly utilizes contextual correlativity to reduce stroke.Present phonetics input method need provide two kinds of information could import a Chinese character to computing machine: first, syllable information, need keystroke (to encode for average 3.5 times and 2.1 times respectively with spelling or simplicity input syllable for spelling, because a pinyin string may be another substring, therefore need when this string end of input, to add a space and distinguished, just on average have more keystroke 0.4 time); The information of which word in numerous unisonance Chinese characters of the second, one sound correspondence is by the numerical key input at present, provides this information on average to need keystroke 1.4 times (consideration frequency).Therefore, average keystroke of word of the every input of phonetics input method is between 3.5-4.9 time, and it is compared with the theoretic 13 times limit very big redundancy, illustrates that wherein some information is omissible.Find that after deliberation second category information above-mentioned is omissible.The fact shows that the people just can know its content after hearing pronunciation in short, and ambiguity can not take place, if the pronunciation of the known a word of this explanation, its contained Chinese character is can be well-determined.Although see isolatedly, the corresponding a lot of Chinese characters of pronunciation in the specific context environment, have only a kind of selection, that is to say, if the context of investigating is abundant, just can be mapped pronunciation sequence of Chinese (pinyin string sequence) and Chinese character one by one.This method is just in view of this point, and a sentence is done as a wholely to consider, comes unique Chinese character of determining each phonetic correspondence by context, and saved this link of manual word selection.Stroke can reduce so on the one hand, the more important thing is owing to accomplished that really the speed of input is just faster with phonetic touch system input Chinese character.The core technology of said method is the automatic conversion of phonetic to Chinese character.Can use based on the sentence comprehension method of corpus statistics and finish this conversion.Phonetic can be seen as a correspondence problem to the conversion of Chinese character, is exactly the pronunciation S=(S of a known sentence S ₁, S ₂..., S _N) finding out should corresponding which type of Chinese character speech string W=(W ₁, W ₂..., W _N), (in short always can be divided into several speech, comprise a words) is according to maximum posteriori criterion: W=ArgMaxP (W ^(j)/ S) (1) W ^(j)By some candidate's sentences (word sequence) of input sentence.According to the Bayes formula, and the independence of P (S) and J, have: W=ArgMax{P (S/W ^(j)) P (W ^(j))=ArgMax{P (S ₁, S ₂..., S _N/ W ₁, W ₂... W _M) P (W ₁, W ₂..., W _M) (2) according to markov hypothesis and independent output hypothesis, have: $P({\mathrm{<figure-callout\; id="1"\; label="W"\; filenames="C9710195100351.png"\; state="\{\{state\}\}">W</figure-callout>}}_1,{\mathrm{<figure-callout\; id="2"\; label="W"\; filenames="C9710195100351.png"\; state="\{\{state\}\}">W</figure-callout>}}_2,...,W_M)=\underset{i=1}{\overset{M}{\mathrm{\&Pi;}}}P(W_i/W_{i-1})$ P(S ₁，S ₂，…，S _N/W ₁，W ₂，…，W _M) $=P({\mathrm{<figure-callout\; id="1"\; label="S"\; filenames="C9710195100351.png"\; state="\{\{state\}\}">S</figure-callout>}}_1,...,{\mathrm{<figure-callout\; id="1"\; label="S\; n"\; filenames="C9710195100351.png"\; state="\{\{state\}\}">S</figure-callout>}}_{n_{}}/W_1)P({\mathrm{<figure-callout\; id="1"\; label="S\; n"\; filenames="C9710195100351.png"\; state="\{\{state\}\}">S</figure-callout>}}_{n_{}},...,{\mathrm{<figure-callout\; id="2"\; label="S\; n"\; filenames="C9710195100351.png"\; state="\{\{state\}\}">S</figure-callout>}}_{n_{}}/W_2)...P(S_{n_{m-1}+1},...,S_N/W_M)---\left(3\right)$ Wherein,

Corresponding W _K+1Pronunciation, do not causing when obscuring that in order to write conveniently, candidate's sequence number J has been omitted.After the Chinese character of multitone is regarded as several different words: $P(S_{n_k+1},S_{n_k+2},...S_{n_{k+1}}/W_{k+1})$

Therefore, the only remaining formula (2) of asking of the problem of calculating formula (1), (3) and (4), it is to obtain by the statistics to existing a large amount of articles (corpus).When realizing, count the value (seeing [5,6]) in formula (3) and (4) in advance, then the pinyin string of input is calculated a most probable Chinese character sentence, computation process is finished automatically by computing machine.Because it is to be foundation with the probability that this method is selected Chinese character, therefore certain mistake is arranged unavoidably, but error rate can not surpass 5%.Select if add tone behind input Pinyin, error rate is no more than 2%.The word not right to automatic understanding also can carry out manual intervention.Simultaneously, this method allows User Defined vocabulary, use self-defining vocabulary after, error rate can also reduce.Above error rate is as the criterion with the various articles on the input newspaper, and the testing material total number of word surpasses 1,500,000 words.The intelligence phonetic letter input method is utilized the contextual correlativity of natural language, has realized the automatic conversion of phonetic to Chinese character, has realized with phonetic touch system input Chinese character.Average stroke of Chinese character of intelligence spelling input method input is 3.5, near present various quick input methods; The average stroke of intelligence simplicity only is 2.11, and is all faster than present various input methods.This method meets the custom that people use spoken and written languages, need not remember loaded down with trivial details code table, is easy to grasp.In use, needn't consider to divine by means of characters, can realize the input of completing.Intelligent input mode is then more preceding goes a step further for the whole sentence of intelligent phoneme-shape code input method among the present invention, its average stroke lacks 1.0-2.0 time than intelligent spelling input method, and the comfort level that it uses is well more a lot of than intelligent simplicity, and is almost similar with intelligent spelling.Therefore, comprehensive it be better input method.Therefore whole sentence intelligent input method can avoid importing wrongly written or mispronounced characters to a certain extent owing to be to be based upon on the basis of natural language understanding.The work of this respect also has a lot.With regard to the intelligent input mode of whole sentence of intelligence phonetic letter and intelligent phoneme-shape code, be that the coding of hypothesis input is entirely true at present, also can be improved to from now on when allowing certain mistake input, still can export correct Chinese character.

The intelligent input mode of whole sentence in the intelligent phoneme-shape code input method of the present invention is on the basis of whole intelligent understanding technology of sentence and auxiliary check and correction technology, for the various encoding schemes of individual character input mode or only import the multiple encoding scheme of first or preceding several codes or the like of each Chinese character basic announcement sign indicating number, with Markov-chain model of introducing previously and whole sentence intelligence input principle obtain satisfy each condition and the big Chinese sentence of possibility appears as output, provide some possibilities time maximum sentence to be selected simultaneously, and may provide various marks in bigger place to makeing mistakes, so that further check and correction is revised.The intelligent input mode of whole sentence of intelligent phoneme-shape code can show the Chinese character words and phrases in the user imports the process of each Chinese character of sentence in real time.The user needn't wait the phonetic of the whole sentence of being totally lost, and just can see the Chinese character of input.Just in case strike wrong key or phonetic mistake, also can in time find, revise easily.The user also needn't remove to set up the dictionary of oneself specially, and in the input Chinese character, system can set up and safeguard automatically dictionary, the specialty characteristics of adaptive user automatically.Whole sentence intelligent input technique and integrating words and phrases technology combine, and can increase input, and the correctness of check and correction and identification also can improve the identification computing velocity, increases real-time and Practical Intelligent function.This intelligent input method can also have automatic memory, self study, revise dictionary adaptively to adapt to user's characteristics, select the error category be used to proofread adaptively, select adaptively and the multiple intelligent function of the type matrix character library used when proofreading and correct the identification Chinese character and specialty characteristics of adaptive user or the like.The intelligence of this system is estimated more and more higher like this.Above-mentioned whole sentence intelligent input method is imported in intelligence, the literal check and correction, Chinese Character Recognition, the Chinese speech input, replace digital postcode with mailing address letter postcode, replace digit phone number with alphabetical telephone number, and function or the like many computword inputs field of realizing the Chinese display beeper on the English digital beeper also has and variously uses widely.(see list of references [7.] for details.)

Obviously, the principle of whole sentence intelligent input method can be used for realizing to phonetically similar word or easily mix the function that the nearly sound word of fallibility assist literal to proofread.The nearly sound word of word that an important feature of Chinese is exactly a unisonance and easily mixed fallibility is many especially.Therefore, the auxiliary literal check and correction function of this respect is with regard to particular importance.

In addition, the errors in text of finding in the literal check and correction has words to obscure mistake, grammar mistake and other mistake.Wherein, grammar mistake can be checked check and correction with syntax rule, mark make mistakes may be bigger the place, and provide the correct sentence of usefulness for reference.Now, the grammar mistake part does not also find the method for the whole sentence intelligence of effective application input principle, can only proofread with syntax rule.Obscure mistake for words, then can set up the dictionary that all words that the easy mistake of each Chinese character by words write as are formed.When the check and correction article, obtain the probability that proofread original text sentence occurs with the described whole sentence intelligence input principle of first-half, a few Chinese character in this sentence by the Chinese character in the aforementioned dictionary substitute back (for example: substitute being restricted to of Chinese character and only substitute a word) institute might in the probability that occurs of the sentence of possibility maximum and this sentence more but be no more than sentence half several Chinese characters by the Chinese character in the aforementioned dictionary substitute back (for example: substitute being restricted to of Chinese character and substitute more than a word but be no more than half of sentence) the probability of sentence appearance of possible middle possibility maximum.If second and the 3rd probability of first likelihood ratio is all big, think that then this sentence does not have words to obscure mistake.If second and the 3rd probability of first likelihood ratio is all little a lot, (for example: 0.8 times than the two is all little) thinks that then this sentence has words to obscure mistake.At this moment, can make mark to words replaced in this sentence or part, and occur the correct sentence of the sentence of possibility maximum after will substituting as usefulness for reference.If situation is not then judged between between the two.If second probability of first likelihood ratio is big but littler than the 3rd probability, then consider the ratio of first probability and the 3rd probability.If this ratio is very little, (for example: this ratio limits 0.5 less than estimating) thinks that then this sentence has words to obscure mistake.At this moment, can with above handle equally when having words to obscure mistake.Otherwise, think that then this sentence does not have words to obscure mistake.Substitute the restriction of Chinese character in the time of can constantly suitably adjusting this estimation limit and calculate second probability and the 3rd probability in the original text sentence, make this proofreading method accuracy higher, and look after other each side factor better.For example: to " appointing people's long live! " when proofreading, calculate with said method, can obtain second little result of probability of first likelihood ratio.Therefore, system thinks that this sentence has words to obscure mistake, and " appointing " word is wherein carried out mark, provides correct sentence " people's long live of usefulness for reference simultaneously! ".This method combines with the integrating words and phrases technology, can improve computing velocity, increases real-time and real-time intelligent function.This method can also have automatic memory, self study, and the adaptive user requirement selects the error category be used to proofread to handle specialty characteristics with adaptive user or the like multiple intelligent function with situation respectively adaptively.

Error category dictionary in the said method can have unisonance fallibility dictionary, nearly sound fallibility dictionary, dialect fallibility dictionary, fallibility dictionary familiar in shape, non-standard Chinese word library, all-phonetic input method fallibility dictionary, double-spelling Chinese character input method fallibility dictionary, five character-shape input method fallibility dictionaries, natural code input method fallibility dictionary, sieve code inputting method fallibility dictionary, intelligent phoneme-shape code input method fallibility dictionary among the present invention, various Chinese character recognition software fallibility dictionaries, the various integrated dictionary of above-mentioned various dictionaries ... or the like multiple dictionary.The dictionary of these different error categories can be used for beating using to listen, see and beat, want the mode of beating and use all-phonetic input method, double-spelling Chinese character input method, five character-shape input methods, the natural code input method, sieve code inputting method, input methods such as the intelligent phoneme-shape code input method among the present invention, and use different Chinese character recognition softwares to carry out output file that Chinese Character Recognition generates and the Chinese character recognition software that can't determine input method or use ... or the like multiple situation proofread.For various different situations, can determine a basic dictionary and corresponding check and correction parameter earlier.Situation about taking place when proofreading according to reality is later on constantly carried out self study by system.If a right mistake is looked for by system, just obtain a positive increment.If system has confused a mistake, just obtain an anti-increment.If system has missed a mistake, also be equivalent to confuse a mistake, also obtain an anti-increment.At this moment system can automatically increase, and reduces or revises corresponding error classification dictionary and check and correction parameter, estimates so that check and correction software constantly increases reliability and intelligence in the process of self study.In a word, whole sentence intelligent input method has many important application in literal check and correction or the like many computword inputs field among the present invention.Chinese Character Recognition, particularly handwritten Kanji recognition are the key subjects in the computword input field.At present, the greatest difficulty that it runs into is that so-called bottleneck problem is how further to improve the success ratio and the reliability of Handwritten Chinese Character Recognition, to reach the purpose of practicability.At present, for the handwritten Chinese character of standard, discrimination still can reach 60%-95%, but for nonstandard handwritten Chinese character, discrimination is still very low.Present aggregate level does not also reach the purpose of complete practicability.For this reason, the present invention proposes a kind ofly can improve the handwritten Kanji recognition rate greatly on whole sentence intelligent input technique and auxiliary literal check and correction technical foundation, makes it to reach rapidly the recognition methods of practicability level, promptly whole sentence intelligence check and correction recognition methods.To each Chinese character in the sentence,, then need not use following householder method if can identify all Chinese characters exactly according to general recognition methods at present.Otherwise if, then to those Chinese characters that can't accurately identify, determine one group of Chinese character that the Chinese character in its font image and the original text is similar to according to general recognition methods at present, add that with these Chinese characters those Chinese characters that accurately identify may form sentence according to the original text order with any respectively then.Consider two factors, the possibility that (1) these sentences occur, the summation of the corresponding font image similarity degree that is identified font in (2) each approximate Chinese character and the original text.Be used in the overall target of calculating on these two factor bases, obtain one might in the best sentence as output, and provide the sentence to be selected of some suboptimums, may demonstrate certain mark in bigger place, so that further check and correction is revised to makeing mistakes.So just can improve present handwritten Kanji recognition rate greatly.Can constantly suitably adjust the various parameters in the said method, similar character library and algorithm make the discrimination of this method higher, and can comprehensively look after other various factors.For example: when using existing recognition methods, " people's long live " might be identified as " going into people's long live " by computation error.But during the recognition methods among use the present invention, computing machine is found out first Chinese character automatically may can be identified as two words, be respectively " people " and " going into ", take all factors into consideration aforesaid two factors (1) and (2) then, possibility to occur much bigger for the sentence of correspondence when discovery was identified as " people " word, and font is more or less the same.Therefore, be identified as " people " and be used as optimal selection, export " people's long live ", and provide the sentence to be selected of some suboptimums, as: " going into people's long live " for " people " word and " going into " word, will provide suitable mark.To improve discrimination and recognition methods like this.This technology can also combine with the integrating words and phrases technology, can improve computing velocity, increases real-time and real-time intelligent function.This method can also have automatic memory, self study, the requirement of adaptive user is selected to be used to proofread the error category of usefulness adaptively, selects adaptively and the many intelligent functions of the type matrix character library used when proofreading and correct the identification Chinese character and specialty characteristics of adaptive user or the like.In the application aspect on-line handwritten Chinese character identification, this method can also have and word for word shows Chinese character in real time, and constantly utilizes the contextual information real time modifying to proofread and correct or the like multiple intelligent function.

In fact said method is the whole intelligent understanding technology of sentence of a kind of application of Chinese character recognition software and the post-processing technology of auxiliary literal check and correction technology.Can all note with Chinese character inequality that identified originally and the Chinese character that identified originally proofread and correct the back through said method.After last automatic check and correction is finished, edit through artificial check and correction again, determine correct text.Therefrom can find out correct but the Chinese character that the aftertreatment check and correction is wrong on the contrary of original identification learns as anti-increment.Find out original identification error simultaneously but the correct Chinese character of aftertreatment check and correction is learnt as positive increment.In this self study process, can constantly increase, reduce or the similar character character library that will use when revising the aftertreatment check and correction in corresponding Chinese character and constantly revise various aftertreatments check and correction parameters.The reliability and the intelligence that can improve constantly this software are like this estimated.Also can develop a kind of concrete Chinese character recognition software that do not rely on, and to aftertreatment check and correction software all or that a part of Chinese character recognition software is suitable for.Can also develop a kind of internal information of not using various Chinese character recognition softwares, and as long as directly software is proofreaded in the aftertreatment that the output file of various Chinese character recognition softwares is proofreaded.In fact this is a special example of foregoing check and correction software.This software is more easily developed.

Obviously, the whole sentence intelligent input technique among the present invention also can be applied to other Chinese Character Recognition fields.In a word, the whole sentence intelligent input technique among the present invention is in Chinese Character Recognition, particularly handwritten Kanji recognition, or the like many computword inputs field many important use are arranged.

Phonetic entry is key subjects in the computword input field.Now a main difficult problem in the Chinese speech input field is the nearly sound word that a large amount of phonetically similar words is arranged in the Chinese and easily use with, and which Chinese character only can't determine what will import with pronunciation inputting method is.Therefore, the Chinese speech input technology can't reach the level of practicability for a long time.

Now, utilize described whole sentence intelligent input technique of former joints and auxiliary literal check and correction technology, the nearly sound character library of some unisonances that the sound work to be identified that will use in the time of can being created as all phonetically similar words and the nearly sound word of easily using with by phonetic entry indicates.When phonetic entry, find out the corresponding nearly sound Hanzi font library of unisonance of each pronunciation in the said sentence of speaker earlier.Then for whole word, utilize methods such as whole sentence intelligent input technique and auxiliary literal check and correction technology, obtain the possibility maximum appears in the different Chinese character combination in any in each character library and a Chinese sentence as output, and provide the sentence to be selected of some suboptimums, may big part provide mark for makeing mistakes, proofread modification so that further use playback to make the mutual method of selecting by the user.This technology can also combine with the integrating words and phrases technology, can improve computing velocity, increases real-time and real-time intelligent function.This method can also have automatic memory, self study, the many intelligent functions of adaptive user requirement or the like.

This phonetic entry technology can combine with the telephone code technology described in the chapter 4, directly by phone with above-mentioned phonetic entry technology to the computing machine input characters, particularly input is Chinese.Run into and to proofread when revising, can use playback to ask the user to select and proofread the mode of modification alternately.Can use this technology like this and set up a kind of foregoing suitable China's actual conditions, convenient and cheap, the Internet system that can conduct interviews by phone.

In a word, the whole sentence intelligent input method among the present invention has many important application in Chinese speech input or the like many computword inputs field.Above-mentioned these technology will be brought a series of revolutionary progress for many computwords input field and compuphone service field.

Because identification handwritten numeral code will be distinguished ten numerals of 0-9, discern hand-written alphanumeric codes and will distinguish 26 letters.Therefore, the identification difficulty of the two, discrimination and identification certainty are more or less the same, and are at least on the same order of magnitude.At present, each post office of China has used a computer and has discerned the postcode of writing with the handwritten numeral code automatically, carries out the automatic go-on-go of mail with computing machine simultaneously.Handwritten numeral code discrimination and the identification certainty of using this technology to obtain can meet the demands.Now, along with the develop rapidly of computword recognition technology, technical indicators such as the discrimination of hand-written alphanumeric codes and identification certainty also can meet the demands substantially.For hand-written alphanumeric codes, also accomplishing uses a computer in the post office discerns and the automatic go-on-go of mail automatically, and discrimination and identification certainty are more or less the same.So just might replace postcode with hand-written alphanumeric codes.When using postcode now, the sender does not often know or forgets receiver's postcode.It is also very inconvenient to search postcode, nor may make greatly the digital postcode of all addresses can both be in the post office or certain unit find.Use hand-written alphabetical code book to write mailing address and replace postcode just to solve this difficult problem, greatly facilitate the user.If the Chinese character input method among use the present invention is write the address, can significantly reduce alphabetical code length, and convenient and practical, easily learn easily note.Above-mentioned technology can be used in combination with whole sentence intelligent input technique and address common phrase integrating words and phrases technology, can increase the correctness of identification, improve the identification computing velocity, improve real-time, increase the Practical Intelligent function.And automatic memory can be arranged, the many intelligent functions of self study and adaptive user characteristics or the like.When the sender knows or is ready to ask about receiver's digital postcode, still can use original digital postcode.Originally all ways about digital postcode still remain unchanged.Just the post office increases some projects for user's service.For example: the sender can not post a letter when knowing receiver's digital postcode yet, just will use the higher special envelope of charge.Can sell two kinds of envelopes.A kind of comparatively cheap, but require the user to not be afraid of trouble, as must to use the rotated type alphameric type matrix that prints off the rotation number word code type matrix of projection time on a kind of similar film ticket of selling at the cinema to make printer.The user can arrive in post office and use the printer that makes with rotated type alphameric type matrix in the post office or oneself buy above-mentioned printer.If often post a letter like this, buy above-mentioned printer and also can receive economically.When the user uses this comparatively cheap envelope, can use the alphabetical postcode that above-mentioned printer prints off to be needed on envelope.Like this, the font of printing off is exactly standard printer's letter and numeral, and has only a kind of standard letter.It is lower to discern difficulty like this, and discrimination and identification certainty are all than higher.Reliability than identification handwritten numeral code is taller a lot.So promptly solve the difficulty problem of identification fully, solved user's difficulty again.Another can charge higher.The user can buy back, and uses stroke handwriting letter postcode on envelope.So promptly made things convenient for the user, had an economic benefit again.Higher to the charge of this envelope, can guarantee that the service item that increases has than high yield, and the part of unnecessary income can be used for increase equipment and improvement technology that it is general to make this technology can be step by step develop into the whole nation from one or two experimental city.Can oneself support oneself like this, need not country spend too many money.The user is in order to post a letter, be reluctant that again spended time or energy go to search digital postcode, can take a little energy more and arrive in post office and use the alphabetical coded word mould of rotation, perhaps spend more a little money and buy an alphabetical coded word mould of rotation, perhaps spend more a little money and use second kind of envelope to post a letter.Can be marked with different block letter marks at the privileged site of two kinds of special envelopes.For example: Y-block letter and S-handwritten form.During the identification of the automatic go-on-go of computing machine, check mark at first.If unmarked, discriminating digit postcode in digital zip box if nil postcode or digital postcode are not right, is pressed handwritten form or block letter identification letter postcode more earlier.If identify, can be delivered.But, when unmarked, can be judged as non-special envelope in order to guarantee income.If at this moment nil postcode or digital postcode are not right, can refuse to know, discern and judge by the people.If the special envelope of right and wrong can be delivered, can not deliver yet, return the original place.Owing to do not guarantee to deliver the non-special envelope that does not have digital postcode, the user generally is reluctant the risk of emitting mail to be return again, so the user is by the special envelope of boost.If be labeled as Y, then discern according to block letter.If the user does not use block letter to print off alphabetical postcode being labeled as on the special envelope of Y, then to emit mail by the risk of knowing and miscarrying by mistake, responsibility is thought highly of oneself by the user.If be labeled as S, then discern according to handwritten form.Can suitably heighten and refuse to know limit, improve reliability.The mail of refusing to know can be discerned with manual method, improves reliability.Because this special envelope charge is higher, such manual service needs and is worth.This envelope can be bought back home, and the user can use stroke handwriting letter postcode, and sends mail near the mailbox the dwelling.At this moment, need not write digital postcode and replace, also can post a letter with hand-written alphabetical postcode.If at this moment need not mark the special envelope of S, mail may be return.If use the special envelope of mark S, then mail can guarantee to send it to.Above-mentioned alphabetical postcode can be used and simplify a kind of mode that draws according to plain mode in this input method and encode.It can (promptly be economized by the big zone in the address, city, autonomous regions etc.) initial of preceding two the word Chinese phonetic alphabet of Chinese title adds that a repeated code rank-numeral is added zonule (perhaps address and the unit) initial of each Chinese characters phonetic of Chinese full name and last repeated code rank-numeral is formed.Wherein, the repeated code in the big zone name seldom and is remembered necessary.Them also will be used in other address of areal.For example: economize, municipality directly under the Central Government and the essentially no repeated code of autonomous region's one-level have only Shanxi Province and Shaanxi Province, can represent that Hebei province and Hubei Province can represent with HB1 and HB2 person HEB and HUB with SX1 and SX2, Henan Province and Hunan Province can represent with HN1 and HN2 or HEN and HUN.In zonule or littler three grades of little address units, may have some repeated codes, but generally can not surpass 10, at most also can not surpass 100, two repeated code rank-numerals of so maximum needs.As long as remember above-mentioned three groups big area-name repeated codes and some address, zonule name repeated codes, just can use this alphabetical postcode.Do not have the repeated code part for major part, directly write and get final product, need not to search and remember.For the part that repeated code is arranged, then to search and remember.If but generally remember the repeated code sequence number, as long as promptly remember the 1-3 bit digital.Can reduce many memory capacitances like this.For example: address " Shanxi province Taiyuan city waterworks " can be represented with " SX1TYSZLSC ", and address " in Yongan, Chaoyang District, Beijing City " can be represented with " BJ1CYQYAL ".Wherein, the repeated code sequence number is arranged since 0 or 1 according to frequency of utilization.This respect relatively thing of difficulty is to be difficult to know whether repeated code in advance, and how many repeated code sequence numbers is.The general alphabetical postcode that in various advertisements, all will write digital postcode and have the repeated code sequence number.If do not write alphabetical postcode, the user can think not have repeated code.If mail wastes time because repeated code is arranged, responsibility is thought highly of oneself by the producer of advertising.If the user is worried, can use the input in Chinese telephone code technology of introducing in the chapter 4 to inquire about to relevant postcode phone automated inquiry system.As long as after having selected function corresponding number, input user and the corresponding input in Chinese telephone code of alphabetical postcode that will write on the envelope, can quote automatically in the phone all with yard address, digital postcode and alphabetical postcodes.The user can check whether the alphabetical postcode of oneself importing is correct.Both can use individual digit to represent the telephone code technology of a letter, also can use the telephone code technology of a letter of two numerals, can certainly the two use together.Digital postcode and alphabetical postcode can be quoted together in the phone.The user generally can use digital postcode, if but the user does not remember digital postcode, and perhaps worried to only using a kind of postcode, then can use alphabetical postcode.So just solved this difficulty of postcode that is difficult to search arbitrary address.When the user is unwilling to search postcode with phone or before the actual use of this system, for the repeated code situation of big zone and middle zone name in the address, the pamphlet that the user can buy all these repeated code situations of record solves this problem.Such repeated code situation at most only has about hundreds of, and pamphlet can be too not thick, necessarily can load.The user also can arrive in post office and search pamphlet or inquire about with relevant computer software, can also utilize aforesaid postcode phone automated inquiry system to inquire about.Like this, the coincident code problem of big zone in the address and middle zone name can solve substantially.If recognition system runs into the big zone in the address and the ambiguity situation of middle zone name, generally refused to know.Discern or return the original place again by manual method.For other repeated code situation, because situation is too many, it is then too thick to weave into handbook.Can not allow a large number of users all have, search also very inconvenient.Therefore run into repeated code situation system and will refuse to know, replenish by manual method and handle.Like this, because use the situation of alphabetical postcode to want much less than the situation of using digital postcode, how many total labor workload can not increase.The average labor workload of handling every envelope letter postcode mail at most also only can increase several times than the average labor workload of handling every envelope numeral postcode mail.Because the former charges higher, doing like this still is what be worth.

In a word, above-mentioned technology can be improved the automatization level and the service level of China's post industry greatly, and can be hopeful to open the system of the postcode of using the Help by Phone arbitrary address.Service be can improve, the users and the people made things convenient for.This technology can also be widely used in various surveys, product investigation, the filling in and the many aspects of computer data typing or the like of guarantee statement and various forms.

Similarly, also can replace digit phone number, realize to notify other people when a kind of user changes telephone number, and other people still can put through the switch system of this user's phone with alphabetical telephone code telephone number.Thereby need not often remember telephone number, also can notify other people when changing telephone number.Like this, this system uses very convenient and practical.

Foregoing telephone code technology can be applied to 114 directory enquiry unmanned automated management systems.Can use the telephone code technology to import on phone by the name of enquiring telephone number (or organization), after mutual the selection, system can find out by the directory enquiry sign indicating number automatically and play corresponding recording and quote by the directory enquiry sign indicating number to the user.So just can save 114 directory enquiry personnel, save manpower, save expense, improve automatization level.Further, can use above-mentioned telephone code technology and realize unattended switch system with name or organization automatic telephone switching.Can in a large amount of band branch exchange machine system of China, at first use this technology, set up a band branch exchange machine system that uses above-mentioned telephone code technology with name or organization automatic telephone switching, replace operator, connect phone automatically according to calling name or organization that the caller imported.For example: be the phone in logical Shoudu Iron and Steel Co two workshops, can put through the Shoudu Iron and Steel Co exchange earlier, with the telephone code code of above-mentioned telephone code technology input " two workshops " three words, system just can be switched to phone Shoudu Iron and Steel Co two workshops automatically then.On this technical foundation, can set up a kind of with alphabetical telephone number replacement digit phone number, the switch system of use name (or organization) automatic telephone switching.Can realize that like this user moves or need not notify other people when changing telephone number, and other people still can put through the advanced function of this user's phone.Can also realize that user's telephone number is secret, connect in limited time, change the hotline number, stay the multiple function of hotline number and voice mail or the like when going on business.Also can produce and a kind ofly can use above-mentioned telephone code technology or directly use the letter key of Chinese pin yin dish to carry out input in Chinese, direct input alphabet telephone number is transferred to the phone of the digit phone number of alphabetical telephone number correspondence then automatically by the electronic installation in the phone.Can a small display screen be installed on this phone, check display screen, revise at any time and error recovery while the user can import.After after confirming, just formally transfer to corresponding digit phone number.Wherein, the alphabetical telephone number of digit phone number correspondence can be by user oneself input decision.This phone must be practical, and can have certain market.Can use computer technology such as call voice card technique on the phone that links together with computing machine, to realize this function, also can produce a kind of comparatively cheap telephone for special use specially.

Can at first use a kind of alphabetical telephone code trunk code telephone number that utilizes input in Chinese telephone code keyboard technique.This novel alphabetical telephone code trunk code telephone number can be divided into two kinds.First kind is novel short code letter telephone code trunk code telephone number, it is made up of three numerals, preceding two numerals are input in Chinese telephone code numerals that this toll telephone office is wanted the initial correspondence of preceding two the word Chinese phonetic alphabet of regional Chinese title, and third digit is the repeated code sequence number.This short code letter telephone code trunk code telephone number can be applied to tens or a hundreds of main cities of China.For example: according to Fig. 3. shown in (I) type input in Chinese telephone code keyboard on the telephone code rule stipulated, the short code of Beijing letter telephone code trunk code telephone number can represent with 251 or 250, and the short code letter telephone code trunk code telephone number of Hohhot City ,Inner Mongolia Autonomous Region then can add that the 3rd repeated code rank-numeral represent with 33.Use like this and remember all very simple and convenient.Second kind is novel long code word base telephone code length way area code telephone number, generally is applied to some areas, zonule in each big zone.Can not use the area of first kind of number generally can both use second kind of number.It is by five or six digital compositions, preceding two numerals are provinces that this toll telephone office is wanted the place, zonule, the city, the input in Chinese telephone code numeral of the initial correspondence of preceding two the word Chinese phonetic alphabet of the big regional Chinese title of autonomous region or the like, the 3rd and fourth digit are the input in Chinese telephone code numerals that this toll telephone office is wanted the initial correspondence of preceding two the word Chinese phonetic alphabet of zonule Chinese title, and the 5th or the 5th and the 6th numeral are the repeated code sequence numbers.Repeated code is less and when being less than 10, the numbers that can use five numerals to form.Repeated code is more and more than 10 but when being less than 100, the numbers that can use six numerals to form.General repeated code can be more than 10, more can be more than 100.Usually, can arrange the repeated code sequence number according to the sum of having a telephone installed in each area, many sequence numbers of having a telephone installed come the front.For example: the long code word base telephone code length of Inner Mongolia Autonomous Region Erenhot City way area code telephone number can add that the repeated code sequence number represents with 6635.This like this toll number is fully can be received no longer than six bit digital.As long as everywhere phone generally uses Fig. 3. shown in (I) type input in Chinese telephone code keyboard and corresponding telephone sign indicating number technology, this alphanumeric toll number uses and remembers can be very simple and convenient.It can save the trouble of many memories.

Suggestion stipulates that artificially the character code of toll number area code correspondence is identical with this area's letter postcode.Just can accomplish this point as long as the regulation principle of the two is consistent.Using like this can be more convenient and practical.

The application that realizes the aspects such as function of Chinese display beeper on the English digital beeper mainly is some encoding schemes of utilizing intelligent phoneme-shape code input method among the present invention, wherein previous or preceding several sign indicating number only got in each Chinese character, the only regular in accordance with regulations several Chinese characters that round in the word of brief information are encoded, be made in the encoding scheme of using on the English digital beeper.(encoding scheme can be referring to rearmost explanation in detail.) like this, the original numeric coding scheme of digital beeper is transformed, obtain the alphanumeric codes encoding scheme that ten numerals of 26 letters and 0-9 are formed.Can use English digital beeper,, also can realize the function of Chinese display beeper with the new old encoding scheme of encoding scheme replacement, practicality very easy to use, simple cheap than the cheap manyfold of Chinese display beeper.Especially, using the English digital beeper to transmit the called person surname, name during information such as brief message and time place, all can use this encoding scheme.As long as each beeper station will send the signal of digital code originally and change the corresponding Chinese phonetic alphabet of transmission and the signal of digital mixed code into.The user of use English digital beeper need not increase equipment or do any change and just can use this technology.At this moment, corresponding Chinese phonetic alphabet code is presented on user's the English digital beeper.The user just can draw corresponding Chinese character information through simple study from these Chinese phonetic alphabet codes, search on the code book or multiple platform and not be used in.Can make the English digital beeper use many functions of similar Chinese display beeper like this, make the two about the same easy to use.This technology is convenient cheap, simple and practical, easily learns easily note, will provide great help for the service level of improving the English digital beeper or the like many aspects.Encoding scheme herein also can be applied to aspects such as the automatic message service of foregoing beeper.At this moment, various beepers (comprising the Chinese display beeper, digital beeper and English digital beeper) can use these encoding schemes.(see foregoing encoding scheme for details.)

The English digital beeper encoding scheme of intelligent phoneme-shape code input method is an encoding scheme of serving the English digital beeper system that can realize Chinese display beeper function.It has multiple coded system.Each beeper station and vast beeper user can decide according to the hobby of oneself and use which kind of coded system.For the surname code, this programme provides 2-code scheme and two kinds of schemes of 3-code scheme.I recommend you to use the 3-code scheme.For other brief term, main place name, main unit and public place of entertainment, restaurant, market, hotel, various relevant public services, each office and other provinces and towns mechanism in this city, stadiums and other, ... or the like, this programme provides many yards classification schemes altogether, many code plans, 4-sign indicating number classification schemes, the 4-code plan, 3-sign indicating number classification schemes, multiple schemes such as 3-code plan.Wherein that above-mentioned all kinds of use occasions are divided into ten classes is as follows for classification schemes: 0-is used for the service signal that transmits when this beeper platform is notified the user and relevant system's circular information, 1-is used for various congratulation terms, term of courtesy and relevant information thereof, 2-is used to ask term and relevant information, 3-is used for intelligence aids, each class noun such as time place, and other information, 4-is used for relevant informations such as main place name, 5-is used for relevant informations such as main unit and public place of entertainment, 6-is used for the restaurant, the market, relevant informations such as hotel, 7-is used for various relevant relevant informations such as public services, 8-is used for relevant informations such as each office and other provinces and towns mechanism in this city, and 9-is used for stadiums and other relevant informations.(can also use alphabetical i herein, u, v or whole 26 letters are as group indication.) I recommend you to use many yards classification schemes.(operation instruction can see specifying of various encoding schemes described below for details in detail.) as: ai: like Chinese mugwort, an: peace, ao: Ao.Ba: crust, white cedar, class, Bao Baobao, be: shellfish, bi: finish, not, and the limit Bian, the guest, bo: uncle is thin, bu: foretell the step.Ca: the storehouse is grey, prosperous Chang Chang, and bavin, Cai, Cao, Chao, ce: car, journey becomes, Chen Chen, ci: the pond is slow, cog: high, from clump, cu: Cui, Chu's storage.Da: reach, party wears, pellet, and de: Deng, di: Di's residence of a high official, fourth, tricky, do: Dong Dong, Dou Dou, du: Du is stifled, section,, many.E: Hubei Province.Fa: square room, model Fan, fe: Feng Feng Fengfeng, take fu: pay Fu Fufufu.Ga: lid, sweet doing, high highland Gao, ge: Ge Ge, Geng Geng, go: the public tribute of palace Gong Gong, collude just, gu: Gu Gu turns round and look at, and pipe is closed, Guo state mistake in osmanthus.Ha: breathe out, Hangzhoupro, the sea, Korea Spro, Hao, he: He He and, black, ho: red great flood, waits thick, hu: Hu Hu recklessly, China's flower, Huang, fiery suddenly.Ji: Ji Jiji records and counts the season Ji, merchant, joint, Jiang Jiang Jiang, Jing Jing well, letter, Jin Jinjin, Jiao, jv (or ju): occupy and bring up.Ka: health is high to be shouldered, triumphant,, and ke: Ke, ko: the hole sky, the bandit, ku: the Kuang condition of rectifying is spacious, wool grass.La: the youth, rely, orchid, Lao Lao, le: cold, thunder, li: Li Lilili is strict, and beam is insulted, and honest and clean the white silk connects, woods Lin, Liao, Liu Liu, lo: dragon is grand, building Lou, lu: Lu Lulu Lu Lu, goldenrain tree, sieve white horse with a black mane, lv (or lu): Lv Lv.Ma: the horse fiber crops, big, wheat is bought, and is full, the hair thatch, me: Meng Meng, plum, mi: rice is rotten, and is bright, Min, seedling Miao, mo: not, try to gain mu: Mu Mu.Na: that, south, ne: can, ni: Ni, Nie, peaceful, year, ox button, nog: farming.Ou: Ou Qu.Pa: Pang, Pan, pe: Peng Peng, Pei, pi: the skin Pi, pu: Pu Pupu pounces on.Qi: neat Qi Qi, strong, the minister in ancient times, money, the national muscial instrument celery is admired, Qiao Qiao, Qiu Qiu Qiuqiu enemy, qu: full powers, qv (or qu): flexing Qu.Ra: slowly, re: appoint, ro: Rong Rong melts army, ru: Ruan, Rui.Sa: Sa, sand, mulberry, Shang Shangshang, match, the mountain is single, splendid Shao, se: She, contain and give birth to rope, gloomy, Shen Shenshen, si: teacher of the executing during the stone history, take charge of this, so: Song Song, longevity head, su: the Su Su of Soviet Union, relax, two, general, Sui, water tax, grandson, rope.Ta: Tang's soup, Tan talks, Tao Tao, te: rise ti: iron, field, to: Tong Tong, tu: be coated with Tu.Wa: king Wang, as if ten thousand intact, we: father-in-law, the luxuriant Wei danger of Wei Wei, Wen Wenwen, wu: Wu 5 military Wu witch crows.Xi: why Xi seat practises, the summer, and thank and separate, Xiang Xiangxiang, Xing Xing admires and washes, suffering, Xiao Xiao, xu: Xue, a surname, Xun, xv (or xu): being permitted Xu must the petty official.Ya: Yang Yangyang, tight Yan Yan face swallow is feted, and Yao Yao the one, ye: leaf, yi: Yi Yiyi, should, the cloudy Yin Dynasty of Yin Yin seal, yo: Yong Yong, outstanding trip has yu: month Yue Yue is happy, and Yuan Yuanyuan is far away, cloud Yun, yv (or yu): in Yu Yuyu.Za: hide, Zhang Zhang kills and carries, Zhai, and Zhan's exhibition accounts for Zhan, Zhao Zhao, ze: once, and Zheng, discriminate zi: money,, zo: ancestor, Zhong Zhong, Zou, week, zu: ancestral, all Zhus of Zhu Zhu, the village, a left side, Zhuo.Two-character surname: cy: the chief of the Xiongnu in Acient China, dm: Duanmu, df: east, dg: Dongguo, gs: Gongsun, gl: Gongliang, each beam (unique two-character surname repeated code), hf (or hp): Huangfu, ng: Nangong, nm: south gate, oy: Ouyang, sg: Shangguan, sk: the minister of public works in ancient china, sm: Sima, st: Situ, xh: Xiahou, xm: west door, zl: Zhongli, zs: Zhongsun, zg: Zhuge.Other: qt: other surnames, wz: outer clansman, wg: the foreigner.And for example: ai: like Chinese mugwort, an: peace, ao: Ao.Ba: crust, bai: white cedar, ban: class, bao: Bao Baobao, bei: shellfish, bi: finish bie: not, and bin: the limit Bian, the guest, bo: uncle is thin, bu: foretell the step.Cag: the storehouse is grey, prosperous Chang Chang, and cai: bavin, Cai, cao: Cao, Chao, ce: car, ceg: journey becomes, cen: Chen Chen, ci: the pond is slow, cog: high, from clump, cui: Cui, cu: Chu's storage.Da: reach dag: party, dai: wear dan: pellet, deg: Deng, di: Di's residence of a high official, dig: fourth, dio: tricky, dog: Dong Dong, dou: Dou Dou, du: Du is stifled, dun: section,, duo: many.E: Hubei Province.Fag: square room, fan: model Fan, feg: Feng Feng Fengfeng, fei: take fu: pay Fu Fufufu.Gai: lid, gao: sweet dried, gao: high highland Gao, ge: Ge Ge, geg: Geng Geng, gog: the public tribute of palace Gong Gong, gou: collude just, gu: Gu Gu turns round and look at, gui: osmanthus, gun: close and manage guo: Guo state mistake.Ha: breathe out hag: Hangzhoupro, hai: the sea, han: Korea Spro, hao: Hao, he: He He and, hei: black, hog: red great flood, hou: waits thick, hu: Hu Hu recklessly, hua: China's flower, hug: Huang, huo: fiery suddenly.Ji: Ji Jiji records and counts the season Ji, jia: merchant, jie: joint, jig: Jiang Jiang Jiang, Jing Jing well, jin: letter, Jin Jinjin, jio: Jiao, jv (orju): occupy and bring up.Kag: health is high shoulders kai: triumphant, kan:, ke: Ke, kog: the hole sky, kou: the bandit, kug: the Kuang condition of rectifying is spacious, kui: wool grass.Lag: youth, lai: rely lan: orchid, lao: Lao Lao, leg: cold, lei: thunder, li: Li Lilili is strict, lig: beam, insult lin: the honest and clean white silk connects, woods Lin, lio: Liao, liu: Liu Liu, log: dragon is grand, lou: building Lou, lu: Lu Lulu Lu Lu, lun: goldenrain tree, luo: sieve white horse with a black mane, lv (or lu): Lv Lv.Ma: the horse fiber crops, mag: big, mai: wheat is bought, man: full, mao: the hair thatch, meg: Meng Meng, mei: plum, mi: rice is rotten, mig: bright, min: Min, mio: seedling Miao, mo: not, mou: try to gain mu: Mu Mu.Na: that, nan: south, neg: can, ni: Ni, nie: Nie, nig: peaceful, nin: year, niu: ox button, nog: farming.Ou: Ou Qu.Pag: Pang, pan: Pan, peg: Peng Peng, pei: Pei, pi: the skin Pi, pu: Pu Pupu pounces on.Qi: neat Qi Qi, qig: strong, the minister in ancient times, qin: money, the national muscial instrument celery is admired, qio: Qiao Qiao, qiu: Qiu Qiu Qiuqiu enemy, qun: full powers, qv (or qu): flexing Qu.Ran: slowly, ren: appoint, rog: Rong Rong melts army, run: Ruan, rui: Rui.Sa: Sa, sand, sag: mulberry, Shang Shangshang, sai: match, san: the mountain is single, sao: splendid Shao, se: She, seg: contain and give birth to rope, sen: gloomy, Shen Shenshen, si: teacher of the executing during the stone history, take charge of this, sog: Song Song, sou: longevity head, su: the Su Su of Soviet Union, relax sug: two, sui: general, Sui, the water tax, sun: grandson, suo: rope.Tag: Tang's soup, tan: Tan talks, tao: Tao Tao, teg: rise tie: iron, tin: field, tog: Tong Tong, tu: be coated with Tu.Wag: king Wang, wan: as if ten thousand intact, weg: father-in-law, the luxuriant Wei danger of wei: Wei Wei, wen: Wen Wenwen, wu: Wu 5 military Wu witch crows.Xi: why Xi seat practises, xia: the summer, and xie: thank and separate, xig: Xiang Xiangxiang, Xing Xing, xin: admire and wash, suffering, xio: Xiao Xiao, xue: Xue, xun: a surname, Xun, xv (or xu): being permitted Xu must the petty official.Yag: Yang Yangyang, yan: tight Yan Yan face swallow is feted yao: Yao Yao the one, ye: leaf, yi: Yi Yiyi, yig: should, the cloudy Yin Dynasty of yin: Yin Yin seal, yog: Yong Yong, you: outstanding trip has, yue: month Yue Yue is happy, and yun: Yuan Yuanyuan is far away, cloud Yun, yv (or yu): in Yu Yuyu.Zag: hide, Zhang Zhang, zai: kill and carry, Zhai, zan: Zhan's exhibition accounts for Zhan, zao Zhao Zhao, zeg: once, and Zheng, zen: discriminate zi: money,, zog: ancestor, Zhong Zhong, zou: Zou, week, zu: ancestral, all Zhus of Zhu Zhu, zug: the village, zuo a: left side, Zhuo.Two-character surname: cy: the chief of the Xiongnu in Acient China, dm: Duanmu, df: east, dg: Dongguo, gs: Gongsun, gl: Gongliang, each beam (unique two-character surname repeated code), hf (or hp): Huangfu, ng: Nangong, nm: south gate, oy: Ouyang, sg: Shangguan, sk: the minister of public works in ancient china, sm: Sima, st: Situ, xh: Xiahou, xm: west door, zl: Zhongli, zs: Zhongsun, zg: Zhuge.Other: qtx: other surnames, wzr: outer clansman, wgr: the foreigner.Several brief term encoding schemes: everything is just fine, and 1wsry please return the 2qhbgs of office

… …

(omit temporarily herein.)

And for example: happy New Year, and 1xnh please answer platform 0qft that everything is just fine that wsry please return the qhbgs of office

… …

(omit temporarily herein.)

Happy New Year, and xnh please answer platform qft list of references: [1.] Feng Zhiwei, the entropy of Chinese character, Modern Chinese quantitative test, the Shanghai education publishing house, pp.267-278.[2.] P.F.Brown, et.al., An Estimate of an Upper Bound for the Entropy of

English，Computational?Linguists，Vol.XX.，pp.31-40.[3.]C.Shannon，Prediction?and?Entropy?of?Printed?English，Bell?Systems

Technical Journal, Vol.30, pp.50-64.[4.] Wu Jun, based on the research and the realization of the Chinese speech understanding method of phonetic, Tsing-Hua University's Master's thesis, tutor: king

Make English, 1993.6.[5.] Wu Jun etc., carry out Chinese speech understanding and the conversion of sound word, the 3rd national man machine language's telecommunications with the method for statistics

The art meeting, in October, 1994.[6.]8.F.Jelinek，Self-Organized?Language?Modeling?for?Speech?Recognition，

ICASSP ＇ 91, pp.450-506,1992.[7.] Wu Jun etc., a kind of input method----intelligence phonetic letter input method based on language understanding, Journal of Chinese Information Processing,

Vol.10, No.2, the second phase in 1992.

Claims (6)

1. intelligent phoneme-shape code Chinese character input method is characterized in that: wherein the encode Chinese characters for computer of corresponding each Chinese character generally is made up of three sound sign indicating numbers and two font codes; The initial consonant part that first sound sign indicating number in three sound sign indicating numbers is the sound sign indicating number, initial by corresponding Chinese characters phonetic initial consonant is formed, the simple or compound vowel of a Chinese syllable part that latter two sound sign indicating number is the sound sign indicating number, first letter and last letter by corresponding Chinese characters phonetic simple or compound vowel of a Chinese syllable are formed, if this simple or compound vowel of a Chinese syllable has only a letter, then the simple or compound vowel of a Chinese syllable of encode Chinese characters for computer sound sign indicating number part is only got a letter, if the Chinese phonetic alphabet of corresponding Chinese character is a zero initial, then the sound sign indicating number part of corresponding encode Chinese characters for computer formed in first letter of this zero initial and last letter, have only one when alphabetical as this zero initial, then the sound sign indicating number of corresponding encode Chinese characters for computer part only is made up of this letter; Two font codes of encode Chinese characters for computer are to take into Chinese character apart monomer word parts, and the combination of radical part or stroke parts is got the Chinese Pin Yin initial of first parts and last parts Chinese character title and formed; When the first stroke or finishing touch when relevant stroke constitutes single character, first letter of getting this single character Chinese phonetic alphabet is as font code, when the first stroke or finishing touch only constitute radical with relevant stroke, first letter of getting the radical title Chinese phonetic alphabet is as font code, when the first stroke or finishing touch and relevant stroke neither constituted single character and also do not constitute radical, first letter of the Chinese phonetic alphabet of getting this stroke Chinese title was as font code; Also use phrase coding to import whole phrase apace in addition, perhaps use the Chinese Pin Yin pseudonym monogram to add that the simple or compound vowel of a Chinese syllable monogram imports the Chinese phonetic alphabet of Chinese character apace; From above-mentioned sound sign indicating number, the code symbols that obtains in font code and the phrase coding is made up of 26 English alphabets and symbol " ü " like this, and wherein " ü " replaces with alphabetical V, and order is pressed the corresponding letter key of coding and can be imported Chinese character on computer keyboard.

2. intelligent phoneme-shape code Chinese character input method according to claim 1 is characterized in that: when taking into Chinese character the combination of several parts apart, it is many preferentially to get stroke, and it is few to get stroke again; If there be first of the font code letter of parts and this word sound sign indicating number alphabetical when identical, the pairing parts of this font code will be got two font codes of the Chinese Pin Yin initial of first parts and last parts Chinese character title as this word by the unit construction after having torn open at last again toward divining by means of characters for a short time; If when second font code that is taken out is identical with first font code, then change two font codes getting first parts and second parts correspondence successively into.

3. intelligent phoneme-shape code Chinese character input method according to claim 1, it is characterized in that: when using the whole phrase of phrase coding input, for two-character word, according to second coding of first coding+the second word of second of+the first word coding of first coding+the second word of first word, or one of second coding dual mode of second coding+the second word of first coding+the first word of first coding+the second word of first word formed the phrase coding; For three words, triliteral Chinese Pin Yin initial is connected successively forms the phrase coding; For four words and the above speech of four words, add that by first three word Chinese Pin Yin initial the last character Chinese Pin Yin initial forms phrase coding.

4. intelligent phoneme-shape code Chinese character input method according to claim 1, it is characterized in that: distinguish that for being difficult in the secondary character library rare Chinese character of word sound uses four font codes, each Chinese character divined by means of characters earlier obtain determining two parts of coding, thereby obtain corresponding two font codes, then parts that again will wherein stroke is more or the order of strokes observed in calligraphy is more forward are divined by means of characters and are obtained latter two corresponding font code, at last two font codes and latter two font code are coupled together the encode Chinese characters for computer of forming four font codes.

5. intelligent phoneme-shape code Chinese character input method according to claim 1 is characterized in that: when using the Chinese Pin Yin pseudonym monogram to add the Chinese phonetic alphabet of simple or compound vowel of a Chinese syllable monogram input Chinese character, the QWERTY keyboard that use a computer adds an additional keyboard; Be printed on ch on wherein additional keyboard the 1st row the 1st key, be printed on sh on the 1st row the 2nd key, be printed on zh on the 1st row the 3rd key, be printed on ai on the 1st row the 4th key, be printed on an on the 1st row the 5th key, be printed on ang on the 1st row the 6th key, be printed on ao on the 1st row the 7th key, be printed on ei on the 1st row the 8th key, be printed on en on the 1st row the 9th key, be printed on eng on the 2nd row the 1st key, be printed on er on the 2nd row the 2nd key, be printed on ia on the 2nd row the 3rd key, be printed on ian on the 2nd row the 4th key, be printed on iang on the 2nd row the 5th key, be printed on iao on the 2nd row the 6th key, be printed on ie on the 2nd row the 7th key, be printed on the 2nd row the 8th key, be printed on ing on the 2nd row the 9th key, be printed on iong on the 3rd row the 1st key, be printed on iu on the 3rd row the 2nd key, be printed on ong on the 3rd row the 3rd key, be printed on ou on the 3rd row the 4th key, be printed on ua on the 3rd row the 5th key, be printed on uai on the 3rd row the 6th key, be printed on uan on the 3rd row the 7th key, be printed on ü an on the 3rd row the 8th key, be printed on uang on the 3rd row the 9th key, be printed on ü e on the 4th row the 1st key, be printed on ui on the 4th row the 2nd key, be printed on un on the 4th row the 3rd key, be printed on ü n on the 4th row the 4th key, be printed on uo on the 4th row the 5th key, be printed on ü on the 4th row the 6th key.

6. intelligent phoneme-shape code Chinese character input method according to claim 1 is characterized in that: for from the sound sign indicating number, the code symbols that obtains in font code and the phrase coding can also be pressed the corresponding numerical key of coding by order and import Chinese character on electric live keyboard; Can on the numerical key 2 of telephone keypad, stamp ABC, stamp DEF on the numerical key 3, stamp GHI on the numerical key 4, stamp JKL on the numerical key 5, stamp MNO on the numerical key 6, stamp PQRS on the numerical key 7, stamp TUV ü on the numerical key 8, stamp WXYZ on the numerical key 9; Perhaps on the numerical key 1 of electric live keyboard, stamp ü UV, stamp ABC on the numerical key 2, stamp DEF on the numerical key 3, stamp GHI on the numerical key 4, stamp JKL on the numerical key 5, stamp MNO on the numerical key 6, stamp PQR on the numerical key 7, numerical key 8 stamps STW, stamps XYZ on the numerical key 9.

Images