基于SVM的股票指数预测

The Stock Index Prediction Based on SVM

摘要: 随着中国经济的飞速发展，越来越多的人加入到股市这个大家庭中来。由于股票市场具有高噪声、不确定等特性，使得股票的价格预测极为困难。而较为准确的预测股票价格，有利于人们的投资。本文选用国泰君安大智慧软件中2007年1月4日至2017年12月29日的沪深300指数中2676个交易日数据作为原始分析数据，通过建立支持向量机模型和ARMA模型进行分析并做出短期预测。实验结果：采用支持向量机模型的预测数据与实际数据的拟合度较高，相对误差控制在4%左右；说明支持向量机模型可以对股票市场做出更准确的价格预测，可以为沪深股票市场股票价格走势的研究提供一些借鉴。

Abstract: With the rapid development of China’s economy, more and more people have joined the big family in the stock market. Because of the high noise and uncertainty of the stock market, it is very difficult to predict the price of the stock. And the more accurate prediction of the stock price is conducive to the investment of people. This paper selects 2676 trading days’ data from Shanghai and Shenzhen 300 index from January 4, 2007 to December 29, 2017 in Guotai Junan intelligence soft-ware as the original analysis data, and builds support vector machine model and ARMA model to analyze and make short-term prediction. Results: the data prediction model of support vector machine and the actual data fitting degree is higher; the relative error is about 4%; the support vector machine model can make more accurate prediction of the price of the stock market, and can provide some reference for the study of Shanghai and Shenzhen stock market stock price.