Research On Numerical Sequence Forecasting Algorithm With Internet Text Sentiment Analysis

With the Internet had become one of the most important communication tools,especially the “Internet +” is changing people’s life at any time,in the information age.At the same time,along with the 4G network gradually mature,users of mobile devices are also growing,more and more Internet users happy through media channels every kind of interactive information,to express their goods and services,social events,opinions and emotions.Due to the Internet wide spread,fast and more users,it is necessary to make the data show explosive growth.It is a collective wisdom of the society,after a long time of development and accumulation,by data mining of the big Internet data analysis,user’s emotional state and emotional expression of social media guide can have the ability to predict many social activities.At present,there are still many problems need to be solved in the research of the prediction algorithm based on sentiment analysis,such as the collection of Internet information,the helpfulness analysis of the text,the selection of prediction model variables and the sensitivity of prediction.In this paper,we develop a data collection,text classification interface for these problems,and propose a single variable and multi variable forecasting model based on the emotional tendency of the credible event information,and forecast the commodity price and the real estate market.The research provides a universal acquisition interface for search engine news data acquisition,and the price data acquisition system based on Scrapy framework and the text helpfulness classification model,which is used to solve the commonality of the collected text data in different fields,dynamic Web pages information collection and the helpfulness of the text.The researcher only needs to provide the key words and the price data acquisition Xpath path according to the request of the interface document.It can collect the text data and the related price data conveniently.Then,the text data can be classify,which can be used to predict the research of the algorithm.Based on the text analysis,this paper proposes a single variable and multi variable price forecasting algorithm based on the emotional tendency of the helpfulness events.Due to the time series prediction,the data samples need to be stable and non-trend,so by subtractive method to fit the data elimination of correlation properties.In the single variable prediction model,the SSA-ARMA model is constructed by combining the affective factors of the helpfulness text,and the optimal number of the regression and the movement of the model are obtained through training.In the experiment,the error of SSA-ARMA is obviously reduced,and the prediction efficiency is better.In order to further analyze the effect of emotional factors,the MSA-VAR multi variable prediction model was constructed to analyze the variables of the real estate stock market,which is obvious effect and better results.The last but not least,based on the results of the algorithm,the application software of mobile device price prediction is realized,which is of high practical value.