Unlock instant, AI-driven research and patent intelligence for your innovation.
User character recognition method in sentence-level Chinese character input method and machine learning system
What is Al technical title?
Al technical title is built by PatSnap Al team. It summarizes the technical point description of the patent document.
A technology of Chinese character input and recognition methods, which is applied in the input/output process of data processing, instruments, electrical digital data processing, etc.
Inactive Publication Date: 2013-07-24
HARBIN INST OF TECH
View PDF0 Cites 0 Cited by
Summary
Abstract
Description
Claims
Application Information
AI Technical Summary
This helps you quickly interpret patents by identifying the three key elements:
Problems solved by technology
Method used
Benefits of technology
Problems solved by technology
[0009] In order to solve the problem in the existing machine learning methods that often requires user intervention to obtain the results required by the user, the present invention proposes a user word recognition method and a machine learning system in the sentence-level Chinese character input method
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more
Image
Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
Click on the blue label to locate the original text in one second.
Reading with bidirectional positioning of images and text.
Smart Image
Examples
Experimental program
Comparison scheme
Effect test
specific Embodiment approach 1
[0042] Specific embodiment one: the user's word recognition method in the Chinese input method described in the present embodiment is:
[0043] For the root c, the probability of the root c appearing in the word combination at the position rp is taken as the word-forming ability IWP(c,rp) of the root c:
[0044] IWP ( c , rp ) = C ( Word ( c , rp ) ) C ( c ) - - - ( 1 )
[0045] Wherein, C(Word(c,rp)) is the number of words that root c occurs with position rp in the corpus used for t...
specific Embodiment approach 2
[0071] Specific embodiment two: the online one-time learning method in the sentence-level input method described in this embodiment, the online learning method is:
[0072] Step 1. Align the output path cRoad[M] and the final candidate path wRoad[N] based on the length, and obtain the aligned output path cRoadA[L] and the final candidate path wRoadA[L]; M, N and L respectively represent the number of words contained in these two paths;
[0073] Step 2, set i=1;
[0074] Step 3. According to the information in the language model, calculate p(cRoadA[i]|cRoadA[i-1]) and p(wRoadA[i]|wRoadA[i-1]), and then use these two values to adopt Maximum a posteriori MAP (Maximum a Posterior) probability method to calculate the user adjustment value C with the largest posterior probability A ;Compare (wRoad[i-1],wRoad[i]) and the corresponding C A Added to the user language model library as a binary element;
[0075] Step 4: Set i=i+1, if i≤L, return to step 3; otherwise, one-time learn...
specific Embodiment approach 3
[0089] Specific Embodiment Three: The machine learning system in the sentence-level Chinese character input described in this embodiment is realized by using the user word recognition method described in Embodiment 1 and the online one-time learning method described in Embodiment 2. The system consists of a user word recognition module and an online one-time learning module, in which:
[0090] The user word identification module is used to identify whether the final output result obtained through user intervention in the sentence-level Chinese character input method is a user word, and encode the word judged as a user word, and then store the user word machine code into the sentence level In the user lexicon of the Chinese character input method;
[0091] The online one-time learning module is used for online one-time learning according to the optimal path output by the sentence-level Chinese character input method and the final path obtained through user intervention when the...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More
PUM
Login to View More
Abstract
The invention provides a user character recognition method in a statement-level Chinese character input method and a machine learning system, relating to the technical field of the machine learning of the Chinese character input. The invention solves the problem of final result acquisition through frequent user intervention existing in a traditional machine learning method. The user character recognition method recognizes user characters by adopting word forming capability in opposite positions as an evaluation criterion, and the learning method is started only when the optimal path outputted by adopting the statement-level Chinese character input method and the final output path are different, acquires a probability value by adopting a probability calculation method based on an N-element grammar and acquires a user regulated value CA by adopting MAP (Maximum A Posteriori), and the regulated value CA and the corresponding characters are stored in a user language model base. The machine learning system is a learning system realized by applying the user character recognition method and the learning method. By adopting the technology of the invention, the user intervention number during inputting can be reduced so that a user is easier to acquire a needed output result.
Description
technical field [0001] The invention relates to a user word recognition method and an online learning method in a machine learning method for Chinese character input. Background technique [0002] The machine learning method in sentence-level Chinese character input can automatically adjust the result of the best Chinese character combination according to the user's input habits, and can be applied to various Chinese character input methods and input systems. [0003] With the continuous progress of natural language processing and artificial intelligence theory, Chinese character input technology has also been improved accordingly, but so far there is no Chinese character input technology that can achieve a perfect conversion state, and each technology has its own shortcomings . It is reflected in the pinyin input method that there is no product that can achieve a 100% conversion rate accuracy, and all of them require user intervention to varying degrees and in different wa...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More
Application Information
Patent Timeline
Application Date:The date an application was filed.
Publication Date:The date a patent or application was officially published.
First Publication Date:The earliest publication date of a patent with the same application number.
Issue Date:Publication date of the patent grant document.
PCT Entry Date:The Entry date of PCT National Phase.
Estimated Expiry Date:The statutory expiry date of a patent right according to the Patent Law, and it is the longest term of protection that the patent right can achieve without the termination of the patent right due to other reasons(Term extension factor has been taken into account ).
Invalid Date:Actual expiry date is based on effective date or publication date of legal transaction data of invalid patent.
Login to View More 
Login to View More