Based On The Baidu Index, The SSE 50 Index Forecasting Study

The rapid growth of the Internet,cloud computing and information industries has opened up the Big Data era,where big data is revolutionizing the world today.Data has infiltrated every area of the industry and business functions today and becomes a vital element of production.With the advent of the Big Data era,big data thinking has increasingly changed the research paradigm of modern finance and continuously expanded new fields of research.In China's securities market,the proportion of ordinary investors who do not possess professional knowledge or lack of professional knowledge accounts for about 70% of the total investors.In the face of more than 2000 stocks in the stock market and market information that is difficult to distinguish between real and fake,Difficult to accurately make investment decisions,but the capital market is "profit-driven",how to accurately forecast the stock up and down and profit is the focus of their attention.From fundamental analysis to technical analysis and now use machine learning algorithms to predict the trend of the stock price,the relevant scholars have never stopped studying the field of stock price prediction.Based on the previous studies,this paper studies how to improve the accuracy of the machine learning algorithm in the stock price trend prediction from two aspects of the algorithm and the model index.Through the research,we find that compared with XGBoost algorithm created by Adaboost and GBDT algorithm optimization As a traditional decision tree,SVM algorithm has higher accuracy.In addition to the technical indicators in addition to the indicators,this article also joined the behavioral finance-related indicators,designed to technical indicators and non-technical indicators combined,the comparison found that the same use of XGBoost algorithm,joined the Baidu Index The model has higher accuracy.