Research Of Text Sentiment Analysis On Objective Function And Strategy Optimization

Text sentiment analysis,also called opinion mining,is the field of study that analyzes people's sentiments and emotions from written language This research implemented the vector representation of text based on the deep learning algorithm of word vector by adopting the statistical language model and using the method of machine learning and reached the goal of automatic sentiment analysis through emotions classification of text vector based on classifier.The main work includes the construction of text sentiment analysis system and the optimization of the system parameters.The main innovative point of the research is the study of optimizing the system.(1)We proposed a new learning algorithm of word vector by optimizing objective function,which is T-Doc2Vec.T-Doc2Vec used the Tonimoto coefficient as the similarity between vectors,which is the extension of the cosine similarity function.The effectiveness of the optimized algorithm was proved by the results of testing experiment on IMDB dataset.(2)According to the short aspect of convergence and global of the standard whale optimization algorithm,we proposed a new Hybrid whale optimization algorithm with improved biomimetic strategy(HBWOA)and the effectiveness of the optimized algorithm was proved by the results of testing experiment of the benchmark functions.The HBWOA,firstly improved biomimetic strategy of the standard whale optimization algorithm by Chaos mapping population initialization and adjusting search strategy adaptively,and then improved the convergence process of the standard whale optimization algorithm by combining particle swarm optimization(PSO)algorithm for the advantages of fast convergence.(3)We optimized the original system for text sentiment analysis by integration of HBWOA and T-Word2Vec algorithm.Firstly,we constructed the new system of text sentiment analysis based on T-Word2Vec algorithm with optimized objective function.And then we designed the parameters optimization of the new system through HBWOA in order to improve the performance of whole system.