Text Sentiment Analysis Based On Ordered Classification

In the Big Data Time,the data that people are exposed to are growing in quantity and dimension.The network has abundant quantitative data and text data.Relative to the quantitative data,text data are major,complex,novel and so on.The emotional analysis of the text in the era of large data in this context continue to develop and attract a large number of researchers attention.The research value and application value are paid more and more attention,and how to make full and effective use of the text data is a huge and meaningful challenge.Text sentiment analysis refers to the representation of text and the selection of its characteristic items.The text analysis allows the com-puter to understand and analyze the human language,and it extracts the features and information from the text data.In this paper,we focus on the ordered classification of the text sentiment analysis.We use the neural network language model word2vec to quantify the text data,and then through the pairwise comparison strategy to do emo-tional classification,then complete text emotion analysis.First of all,in order to make better use of the computer to analyze the text data,we need to quantify the text data.In this paper,we use the neural network language model word2vec to quantify the data.Compared with other methods,the model has the advantages of simpleness,high effi-ciency and easy adjustment.In addition,based on the combination of the other method,such as TF-IDF,LDA.Then optimize the word vector.In numerical experiment,we obtained a better result through this method.On the other hand,this paper focus on the ordered classification problem,and proposes an ordered classification algorithm based on pairwise comparison strategy.It describes how to use the pairwise comparison to transform the ordered classification into disordered classification,how to transform the disordered classification results into ordered classification,and propose the feasible strategies in the coding matrix,sample equilibrium,classifier training,and new sam-ple prediction.Compared with other ordered classification algorithms,our algorithm(PairCode)algorithm has better performance in ACC,MAE and MMAE.Howev-er,due to the characteristics of PairCode algorithm itself,the classification speed is slower.In the follow-up study can be improved.