HMM-based pinyin completion training method, HMM-based pinyin completion model, HMM-based pinyin completion method and HMM-based pinyin completion input method

Abstract

The invention discloses an HMM (Hidden Markov Model)-based pinyin completion training method, an HMM-based pinyin completion model, an HMM-based pinyin completion method and an HMM-based pinyin completion input method, belongs to the field of medical information, and aims at solving the problem of improving the completion speed and accuracy, and the method is characterized by comprising the following steps: S1, acquiring a legal complete single-character pinyin string; S2, obtaining a possible to-be-completed pinyin string corresponding to the legal complete single-character pinyin string; S3,obtaining a corresponding relationship between the to-be-completed pinyin string and the complete pinyin string; S4, obtaining complete pinyin contents corresponding to the training data, wherein thepinyin contents are separated according to individual characters; and S5, performing statistical learning on the training data, and acquiring the initial probability, the emission probability and thetransition probability. The process of traversing all completion results can be avoided, and the matching efficiency can be improved.

Description

technical field [0001] The invention belongs to the field of medical information, and relates to an HMM-based pinyin completion training method, a completion model, a completion method and a completion input method. Background technique [0002] With the progress of hospital informatization construction, doctors inevitably need to input various text content, such as electronic medical records, examination reports, etc. in the office process. The input method is a main entrance for doctors to interact with computers, and its accuracy and applicability have a great impact on doctors' work efficiency. [0003] At present, what most doctors use is the pinyin input method implemented for the general field. At the same time, due to the relatively slow typing speed of most doctors, they tend to obtain the expected Chinese character results with less input when using general-purpose input methods. [0004] A relatively perfect pinyin input method program, when the user inputs part...

AI Technical Summary

Problems solved by technology

[0009] First of all, in the matching process, due to the use of a method similar to fuzzy matching, each matching process needs to traverse all possible completion results and combinations. When the input content is short and the phrase library is large, the matching efficiency will be affected;

[0010] Secondly, since this method screens candidate results based on the frequency of phrases, the candidate results will completely depend on the training data. If there are word segmentation errors in the training data or phrases that do not exist in the training data, this method will not work well. to complete;

[0011] At the same time, when the input incomplete pinyin string is long, it will be difficult to directly match the appropriate candidate results in the phrase library without additional processing of the input string. If the input string is processed and then matched, it will be Adding extra processing will affect the completion efficiency, and because the selection of candidate results depends on the frequency of each phrase without considering the co-occurrence probability between phrases, the accuracy of long pinyin string completion results will be affected. influences

Images