Research And Optimization On The Technology Of Chinese Text Sentiment Analysis

With the rapid development of Internet technology,the number of people using the Internet has increased dramatically,resulting in a rich textual information on the Internet.Quantifying these textual data yields important data values.The Chinese text sentiment analysis is a method of analyzing and researching text information.At present,the main text sentiment analysis methods are semantic understanding and machine learning.From the perspective of statistical machine learning,this paper focuses on the following three aspects:1?Improved feature selection algorithm.Feature selection is an important step in text sentiment classification,which can effectively reduce the number of feature items and reduce the interference of unrelated feature items.In this paper,based on the analysis of information gain algorithm,according to the deficiency of the algorithm,the corresponding calculation factors are designed to improve the classification effect of the feature words selected by the information gain algorithm.Finally,the design experiments verify the performance of the algorithm after the fusion of the calculation factors.2?By learning the theory of S VM,it is found that the training time of the classifier is affected by the number of input data and support vectors.In the construction of hyperplanes,only the support vector plays an important role.When the training data is too large,the training time is affected.K-means clustering algorithm can retain the characteristics of the original data distribution structure.Combine it with SVM algorithm to reduce the training data set and speed up the training time.The selection of penalty factor and kernel parameter of SVM affects the classification result.With the help of genetic algorithm,the optimal parameter combination can be found and good classification effect can be obtained.The above two algorithms are combined with support vector machines to design experiments to verify the performance of the optimization algorithm.3?Based on the above algorithm optimization,build a text sentiment classification model.Through comparative experiments,the model's emotional classification effect was tested.