Addressing blockchain privacy and efficiency challenges in mobile environments: an optimization strategy for lightweight Application of data analysis in the field of investment

Abstract

 As advancing technology makes a viable option to value a stocks and also determine their economic value has become increasingly important.In the global securities market, changes in stock prices can have a significant impact on changes in the global economy. For enterprises, stock price fluctuations reflect their current financial situation, changes in industry supply and demand, and changes in the international macro environment. [1]When people interact the internet information system, large data which is based on this system will be generated by the people. To some extend, bid data can reflect the law and problems for some certain field, and make some predictions with a certain range. In order to analysis the law of stock price, we need to combine the stock itself and the changing market factors so that we can obtain accurate price forecast. When Solving nonlinear and the complex problem, the LSTM model has better performance to process the data[2]. In addition, the CNN model is also one of the most potential model which can process a large amount of data to improve the accuracy of prediction[3]. This paper will select six indicators to conduct establishment of arbitrage pricing model.Trying to optimize the model and find the factors which are more correlated with stock price in these six factors.We first set up a simple linear regression model and then use stepwise function.Secondly, we set up ridge regression and lasso regression and delete some of these factors. Then we randomly select 70percent of the from the compete data, and set up an evaluation index of the computational model. And finally we find in these data, step regression perform best comparing with the ridge regression and lasso regression, and degree of this model for the EBAY can reach over 90 percent.