Single word and multi-word term integrating system and a method thereof

[0001] 1. Field of the Invention

[0002] The present invention provides a single word and multi-word term integrating system and a method thereof for quickly modifying false predictions of input method editors (IME) for Chinese, English or other languages and for improving prediction accuracy.

[0003] 2. Description of the Prior Art

[0004] As mobile devices prevail, it is more often for users to use mobile devices to input texts to desired applications; therefore, various input method editors (IMEs) have been developed to help users improve their input speed.

[0005] For example, many Chinese IMEs are based on pronunciations, such as Zuyin (or Zuying, which uses phonetic symbols), wherein one phonetic symbol corresponds to one key. When the user wants to input a word, he/she can input the phonetic symbols according to phonetic order of the word. The advantages of using phonetic symbols are: the user doesn't need to know the radicals for a specific word and can input the phonetic symbols as long as he/she can pronounce the word, thereby providing an easy and simple input mechanism; however, the disadvantages are: the user cannot input a word which he/she doesn't know how to pronounce, there could be many homophones for each word, making it time consuming in choosing a correct word. Therefore, an IME using first few codes of the Zuyin IME has been proposed to solve the above problems; however, there could be too many word candidates (maybe pages of words) generated during the inputting process if fewer inputted codes are inputted, making it also time consuming in choosing a correct word.

[0006] In order to solve the problem, a new IME based on Pinying (or Pinyin), which is called Intelligent IME, has been proposed, in which the pronunciation of a Chinese word (or character) is corresponded with an English letter with a similar pronunciation; therefore, the user inputs a sentence by using one letter to correspond to one word, while a string of letters represents the sentence. Thereafter, the system automatically determines the sentence by using these inputted letters. However, the system often fails to identify a single word and provides incorrect combination of words since a Chinese sentence is consisted of single words and multi-word terms; therefore, it could take more time for the user to choose the correct words from the word candidates.

[0007] In view of the deficiencies of the prior art techniques, after years of constant researches, the inventor has successfully proposed a single word and multi-word term integrating system and a method thereof in the present invention.

[0008] It is an object of the present invention to provide a single word and multi-word term integrating system and a method thereof, which can predict certain phonetic symbols to be regarded as a single word and not to combine them with succeeding phonetic symbols, thereby reducing the probability of false prediction.

[0009] It is another object of the present invention to provide a single word and multi-word term integrating system and a method thereof, when the system determines certain phonetic symbols to be regarded as a single word, it moves a select cursor to the position between the certain phonetic symbols and its succeeding phonetic symbols to provide a fast way for the user to choose the desired word or phrase.

[0010] In order to achieve the above objects, the present invention discloses a single word and multi-word term integrating system and a method thereof, which mainly comprises an input unit having a keyboard or other software/hardware device worked as a substitute of the keyboard for a user to continuously input phonetic symbols, wherein the phonetic symbols could be first codes of words or complete phonetic symbols of a word, then the phonetic symbols are displayed in a pinying input display unit for the system to use a full sentence prediction unit to find combinations of the phonetic symbols and to predict corresponding words, which can be further combined into a specific phrase or a sentence to be shown in a full sentence prediction result display unit; when the user continuously inputs a string of phonetic symbols or incomplete phonetic symbols, the full sentence prediction unit could make incorrect prediction of the combination of phonetic symbols, at this time a single word prediction unit can be used to forcibly determine the inputted phonetic symbols to be regarded as a single word and not to combine the phonetic symbols of the single word with succeeding phonetic symbols, and then displays the single word in a single word prediction result display unit, while the succeeding phonetic symbols will be used by the full sentence prediction unit and the full sentence prediction result display unit for combining, predicting, and displaying the phonetic symbols; therefore, the user uses the input unit and the pinying input display unit to choose the correct word or phrase until a final sentence is done, wherein the final sentence is sent to a target, which can be an application or a software component for accepting text input. Therefore, the present invention can improve prediction accuracy and enhance input efficiency.

[0016] Please refer to FIG. 1 for a structural view of a single word and multi-word term integrating system and a method thereof, which comprises:

[0017] an input unit 11 for connecting with an input interface of a mobile communication device or other computer devices, the input unit 11 using a keyboard or other software/hardware device worked as a substitute of the keyboard for a user to continuously input pinying or zhuyin codes and display the input result on a pinying input display unit 12;

[0018] a full sentence prediction unit 13 for combining the inputted pinying or zhuyin codes from the input unit 11 into a corresponding word, phrase or sentence; a full sentence prediction result display unit 14 displaying a result of the full sentence prediction unit for the user to verify and choose a correct word, phrase or sentence;

[0019] a single word prediction unit 15, while the full sentence prediction unit could make incorrect prediction of the combination of pinying or zhuying codes, at this time a single word prediction unit can be used to forcibly determine the inputted pinying codes to be regarded as a single word and to make it as a word candidate with high priority and not to combine the pinying codes of the single words with succeeding pinying codes, and then displays the single word in a single word prediction result display unit 16, while the succeeding pinying codes will be used by the full sentence prediction unit 13 and the full sentence prediction result display unit 14 for combining, predicting, and displaying the pinying codes; therefore the user can use the input unit 11 and the pinying input display unit 12 to choose the correct word or phrase until a final sentence is made;

[0020] an output unit 17 displaying the chosen sentence to a target, which can be an application or a software component which can accept text input.

[0021] Please refer to FIG. 2 for a flowchart of a single word and multi-word term integrating method in the present invention, which comprises:

[0022] Step 201: inputting a part of pinying or zhuying codes of a sentence, or full pinying or zhuying codes of a sentence and displaying the inputted pinying or zhuying codes;

[0023] Step 202: finding combinations of the inputted pinying or zhuying codes, which are not yet recognized, to predict a corresponding word, phrase or sentence, and identifying possible word candidates;

[0024] Step 203: displaying a prediction result with word candidates of step 202;

[0025] Step 204: the user determining if the prediction result is correct; if so, then going to step 208, if not, then going to step 205;

[0026] Step 205: the user determining the frontmost symbol which is predicted incorrectly, moving the input mark (cursor) immediately after said frontmost symbol which will be forced to correspond a single word, and displaying single-word candidates for the user to choose a correct single word of said frontmost symbol;

[0027] Step 206: the user determining if the prediction result of the single word is correct; if not, then going back to step 205 to continue moving an input mark (cursor) to the single word or choosing follow-up single word (since a sentence could be consisted of a lot of single words and multi-word terms) and determining if the corresponding sentence is correct, if so, then going to step 207;

[0028] Step 207: using the correct single word chosen by the user to modify the full sentence prediction result into a correct sentence for the user to make final verification;

[0029] Step 208: outputting the inputted word, phrase, or the sentence to a target, such as an application or a software component which can accept text input and ending the process.

[0030] Please refer to FIG. 3A to FIG. 3G for embodiment views of the method for combining a single word and a multi-word term in the present invention;

[0031] When the user wants to get the following Chinese sentence:

[0032] “” (JinTienDeDuShuHueiZhuTi, which means “The subject of the study group for today”), he or she needs to type:

[0033] JTDDSHZT

[0034] The first two symbols “JT” is recognized as “”, but the target single word “” is not recognized.

[0035] In this stage, the candidate words of the rest symbols “DDS . . . ” could be multi-word terms such as: “ . . . ” (DaDuoShu/DaDuShi/DeDao/DaDao), which are not single words. Therefore, the present invention forcibly determines the inputted codes to be regarded as a single word by indicating the starting symbol, “D”, of the series of symbols “DDS” to identify the word that the starting symbol, “D”, represents, such as by moving the cursor to the position between pinying codes “D” and “D”, so the target words and the inputted codes could be read as:

[0036] “”

[0037] JT D←DSHZT

[0038] And the symbol “D” now could be a single word such as: (Da/Duei/Dao/Di . . . ) with De (“”) falling very far behind, then the user can input pinyin symbol “e” to combine with the symbol “D” to limit the candidate words to “” (all pronounced De). Thereafter, the user chooses “” (De) to finish part of the sentence, and the multi-word term “” (JinTien De) is recorded, that is, the prediction result of “JTD” is confirmed by the user and thereby recognized, wherein the operation is shown as:

[0039] “←”

[0040] JTD DSHZT

[0041] Then, as the candidate words of the rest symbols “DSHZT” show: “ . . . ” (DongShiHuei/aoShiHou/DuShu), while “” (DuShuHuei) is not shown in the list; when the candidate words “” (DuShu) are chosen, the sentence is shown as:

[0042] “ ”

[0043] JTDDS HZT

[0044] Then, as the candidate words of the rest symbols “HZT” show: “ . . . ” (HuoZhe/HungZho/HuangZe), at this time, the user could click on a key to forcibly determine the inputted codes of “” to be regarded as a single word and to move the cursor between the pinying codes “H” and “Z”, so the sentence is now shown as:

[0045] “←”

[0046] JTDDS H←ZT

[0047] As candidate words of “H” show single words such as: “ . . . ” (Han/Hai/Hung/Hou, while Huei falling very far behind), now the user can input pinyin symbols “uei” to combine with “H” to limit the candidate words to “” (all pronounced Huei). Thereafter, the user chooses “” (Huei) and the multi-word term “” (DuShuHuei) is recorded, that is, the prediction result of “DSH” is confirmed by the user and thereby recognized, wherein the sentence is now shown as:

[0048] “ ”

[0049] JTDDSH ZT

[0050] Now the candidate words of the rest symbols “ZT” show “” (ZhuTi/ZhaunTai/ZhuanTi); thereafter, the multi-word term “” (ZhuTi) is chosen to finish the text and sentence input wherein the final sentence sent to the target is now shown as : “” (JinTienDeDuShuHueiZhuTi).

[0051] Therefore, whenever a series of phonetic symbols, which can be a portion of the whole inputted symbols, are not recognized as a phase or a sentence, the starting phonetic symbol of the series of phonetic symbols is indicated to identify the word that the starting phonetic symbol represents, and then the rest of the phonetic symbols in the series of phonetic symbols are converted into corresponding words.

[0052] Please refer to FIG. 4A to FIG. 4F for embodiment views of a single word and multi-word term integrating method using phonetic symbols in the present invention; in which:

[0053] When the user wants to input the following sentence:

[0054] “” ( GinTien DeDienShiGiWeiShiu, which means “The TV maintenance for today”)

[0055] (GT DDSGWS)

[0056] The first two symbols “” (GT) have been recognized as “”.

[0057] As candidate words of the rest symbols “” (DDSGWS) show multi-word terms such as: “ . . . ” ( , DiDien/DaDi/DeDao/DingDin) and no single word, at this time the single word identification mechanism is used first to find a single word, a cursor is moved to between the zhuying codes “” (D) and “” (D), then an arrow would appear in the figure to remind the user to make modifications, at this time the input codes are shown as:

[0058] “←” (GinTien De←DienShiGiWeiShiu)

[0059] (GT D DSGWS)

[0060] As candidate words of “” show: “” (, De/Duei/Dao) with “” (, De) falling far behind; at this time the user can input zhuying symbol “” to combine with the “” to limit the candidate words to single words such as “ . . . ” (all pronounced , De); then the user can choose “” (, De) to finish part of the sentence and have “” (, GTD) shown as a multi-word term, that is, the prediction result of “” (GTD) is confirmed by the user; therefore, the sentence is now shown as:

[0061] “” (GinTien DienShiGiWeiShiu)

[0062] (GTD DSGWS)

[0063] As the candidate words of the rest symbols “” show phrases such as: “ . . . ” ( , DienShi/DienShiGi/DaShiGie, while “” (DienShiGi) is not shown), the user chooses “” (, DS); therefore, the sentence is now shown as:

[0064] “” (GinTien DeDienShi GiWeiShiu)

[0065] (GTDDS GWS)

[0066] As the candidate words of the rest symbols “” show: “ . . . ” ( , GiWang/GiaWu/GieWu), the user now can click on a key to forcibly determine the inputted codes of “” to be regarded as a single word and to move the cursor between the zhuying codes “ (G)” and “ (WS)”, the user can see an arrow appear in the figure to remind him/her to make modification, so the sentence is now shown as:

[0067] “←” (GTDDS G←WS)

[0068] (GTDDS G WS)

[0069] As the candidate words of “” show: “ . . . ” ( , Gin/Ging/Gia/Gin) with “” (, Gi) falling far behind, the user can input zhuying symbol “” to combine with “” to limit the candidate words to single words such as “ . . . ” (all pronounced , Gi), then the user can choose “” (, Gi) to finish part of the sentence and have “” (DienShiGi) shown as a multi-word term, that is, the prediction result of “” (DSG) is confirmed by the user and thereby recognized; therefore, the sentence is now shown as:

[0070] “” (GinTien DeDienShiGi WeiShiu)

[0071] (GTDDSG WS)

[0072] As now the candidate words of the rest symbols “” show: “ . . . ” ( , WuShien/ZuangTai/ZuangTi/WeiShiu), the multi-word term “” (, WeiShiu) is chosen to finish the text and sentence input; wherein the final sentence sent to the target is shown as: “” (GinTien DeDienShiGi WeiShiu).

[0073] The present invention provides a single word and multi-word term integrating system and a method thereof, while compared with other prior art techniques, is advantageous in:

[0074] 1. The present invention provides a method for reducing the probability of false prediction due to incorrect combinations of long Pinying or Zhuying codes, thereby improving the text input speed.

[0075] Many changes and modifications in the above described embodiment of the invention can, of course, be carried out without departing from the scope thereof. Accordingly, to promote the progress in science and the useful arts, the invention is disclosed and is intended to be limited only by the scope of the appended claims.