False Memory Research Method Of Chinese Text Proofreading

Information society is a key feature of high-speed, mass, a variety ofinformation transmission in these vast amounts of information in a hug enumber of errors, seriously affecting the quality and efficiency ofinformation transmission.Learn from existing research results, the paper delves into Chinese fortext editing and proofreading error detection methods, elaborated textediting and proofreading error detection key issues, solutions, algorithms,design principles and other aspects. Chinese text editing and proofreadingmodel for error detection and correction of error detecting errors into twoparts, theoretical analysis and experimental results show that the twoseparate error detection and proofreading help resolve two key objectivesissues, namely confusion set size and Statistics time and space complexityof the language model contradiction between the two is the size of thetraining is expected to result in data sparse phenomena and text data errorscaused by sparse illusion.In text editing error detection, combined with a number of previousstudies have been achievements in the detailed analysis of the Chinese textproofreading task facing the particularity of the data sparseness problem,based on sparse data suggested a solution of the three data smoothingstrategy, all of them are the text window shrinks, smooth delay and clusterwords. N-Text proposed a law text window around the automatic detectionscheme, character properties in the treatment of detection based on theanalysis, based on its local significance this context, considering theprobability and the preceding paragraph, the probability of the latter,combined with three kinds of data smoothing strategy advantages, a bettersolution of the model execution time complexity and space complexity ofthe problem. Experimental data validation, before and after the N-Text lawin the error detection mechanism reporting rate, false negative rate andother indicators perform better.In text editing error correction terms, describes the minimum editdistance concept explained confusing set of role, the use of confusing setdifferent character given different weights strategy from correct the error model screened out large weight as the corrections suggested the correctcharacter. Finally, experimental verification comparison with the dark horseand text proofreading system based on the weight evenly distributed EricMays thinking proofreading models, character weights assigneddynamically correct the error performance of the solution is better.To validate the proposed mo del to detect errors and correct mistakesfeasibility and efficiency of the implementation, the paper carried outexperiments to compare. Experimental data show that the basic completionof the expected goal, to achieve the performance of a mo del based on theexisting achievements have significantly improved.