Method and apparatus for generating alternating content in input method

Abstract

The invention discloses a method for generating candidate content in an input method. The method comprises: caching content input by a user to a screen at the last Nth time, wherein N is an integer more than or equal to 1; and combining content input by the user at the time and the cached content to generate the candidate content. By caching the content input by the user to the screen and using the content input by the user to the screen as subsequent input, the method generates reference and limiting condition of the candidate content, so as to improve accuracy of the candidate content. The invention also discloses a device for generating the candidate content in the input method at the same time.

Description

technical field [0001] The invention relates to the technical field of input methods, in particular to a method and device for generating candidate content in the input method. Background technique [0002] At present, computer users need to use input method software to input Chinese characters or other non-English characters into the computer. The input method software can convert the codes input from input devices such as keyboards into Chinese characters or other characters. For the Chinese pinyin input method, since there are many homophones, after inputting a single word or the pinyin of a word, the input method software needs to generate multiple candidate words for the user to choose. After the user selects the correct word or word on the screen, continue Enter the next word or words. [0003] It is relatively simple to generate candidate content for words and single characters. Generally, they are arranged in order according to the size of the word frequency. The wo...

AI Technical Summary

Problems solved by technology

In the existing input method, the process of generating candidate words for each word or single character input by the user is independent, and there is no mutual relationship between the operations of generating candidate words each time, and the functions of context association and whole sentence generation are only in A single word or word is valid before it is confirmed on the screen, and after the single word or word is on the screen, the input method cannot judge the relationship between the content that has been input by the user on the screen and the content that the user subsequently enters, context association and sentence generation. It doesn't work, and can only generate candidate words simply based on word frequency. For example, if the user wants to enter the sentence "this bedroom is very big", the user's input habit is to enter the sentence twice, and the first time enters "zhejian ", the input method judges that this is a word input, and the first candidate content generated according to the word frequency is "this room". After the user selects "this room" on the screen and continues to input "woshihenda", the input method will generate it based on the existing entire sentence The method generates the candidate sentence "I am very big", but "I am very big" is not what the user wants. At this time, the user needs to manually select other candidate content "bedroom" and "very big" arranged in the back to get " bedroom is big”

And if you input the entire pinyin of "zhejianwoshihenda" at one time, the first candidate sentence generated based on the full sentence prediction based on the maximum probability is "this bedroom is very big". Obviously, when the entire sentence is divided into multiple inputs, the candidate content generation The accuracy rate is not high, and the same effect as the whole sentence generation cannot be achieved. Users need more selection operations to get the correct sentence, which reduces the user's input speed and input experience, and makes the whole sentence generation function useless.

[0009] However, in practical applications, it is basically impossible for users to change their input habits. Therefore, under the existing user input habits, that is, the habit of inputting pinyin in units of individual characters or words, the existing input method cannot realize the complete sentence. The generated effect, the accuracy of the generated candidate content is low

Images