基于机器学习模型的股价波动预测研究 金融市场随着市场经济的不断发展愈发繁荣，其中股票融资占据了很大一部分比例。股票以其高收益的特征吸引了许多投资者，但其同时也具有高风险性，稍有不慎就会给股民带来亏损。经研究发现，股票的特殊性决定了其价格走势很大程度上受投资者心理的影响，投资者的分散性与从众心理使其无法从完全理性的角度看待问题。当前，采用新兴的机器学习方法进行大量数据处理成为一种科学可行的方法，本文通过研究影响股市行情的变量，首先构建爬虫爬取大量股评进行人工分析，之后采用SVM等机器学习算法对数据进行训练，构建情感分类模型，接着构建SVM股价波动预测模型，预测股价波动。研究表明，基于SVM构建的情感分类模型效果最好，通过对照实验发现引入由股评数据计算出的情感指数后模型预测效果会提升，表明股价波动在一定程度上受到股民情绪的影响，将股民情绪指标引入模型对预测股价波动有一定效果。 关键词：情感分类；股价预测；支持向量机；情感指数 Research on Prediction of Stock Price Volatility Based on Machine Learning Model ABSTRACT With the continuous development of the market economy, the financial market has become more and more prosperous, and stock financing accounts for a large proportion. The stocks attract many investors with their high-yield characteristics, but they also have high risks, and a little carelessness will bring losses to shareholders. Studies have found that the particularity of stocks determines that its price trend is largely affected by investor psychology. Investors' dispersion and herd psychology make it impossible to look at the problem from a completely rational perspective. At present, the use of the emerging machine learning method to process a large amount of data has become a scientific and feasible method. By studying the variables that affect the stock market, firstly, the article builds a crawler to crawl a large number of stock reviews for manual analysis. Secondly, the article uses machine learning algorithms such as SVM to train the data to build a sentiment classification model. Thirdly, the article builds a SVM stock price fluctuation prediction model to predict stock price fluctuations. Research has shown that the emotion classification model based on SVM has the best effect. Through the control experiment, it is found that the prediction effect of the model will be improved after the emotion index calculated by the stock evaluation data being introduced. It shows that stock price volatility is affected by the stockholder sentiment in a certain extent, and Introducing stockholder sentiment indicators into the model has a certain effect on predicting stock price fluctuations Keywords：Sentiment classification ,Stock price forecast ,Support Vector Machines, Affective index 目 录 摘要 3 ABSTRACT 4 一、引言 1 （一）研究背景 1 (二)研究意义 1 （三）研究现状 1 （四）研究内容 2 二、相关理论基础 4 （一）方法介绍 4 1.K-最邻近算法 4 2.朴素贝叶斯 4 3.支持向量机算法 4 三、情感分类模型的构建 5 （一）数据采集 5 （二）文本预处理 5 1.人工标注 5 2.中文分词 5 （三）基于机器学习算法的情感分类 7 1.词向量表示 7 3.情感分类模型的构建与评估 8 四、实证以及结果分析 10 （一）基于单纯股票技术指标的股价波动预测 10 1.指标选取 10 2.数据获取 10 3.数据标准化处理 12 4.SVM股价波动预测模型的构建 12 （二）引入投资者情感指数的股价波动预测 13 1. 实时信息与隔夜信息的引入 13 2. 信息量化处理 13 3.应用SVM情感分类模型 13 4.情感指数计算与数据标准化处理 13 5.构建SVM股价波动预测模型并评估 14 五、研究结论与展望 19 附录 20 参考文献 23
*若需了解更多与协助请咨询↓→[电脑QQ][手机QQ]【数据协助】
