基于机器学习的股市拐点影响因素研究

股票市场的成交情况可以充分反映投资者的行为特征并影响整个股市的走势.股票成交明细数据作为股市最底层的交易数据,能够全面地体现股票交易的情况,成为至关重要的股票市场走势判断的参考数据,能够为资本市场监管者在风险监测领域进行决策提供有效帮助.文中提出了一种可以快速地在海量股票交易明细数据中提取投资者交易特征的方法,然后基于逻辑回归、决策树和随机森林等机器学习算法找到股市大盘较大拐点产生的主要影响因素,并预测交易特征变量对股市较大拐点产生的时间范围.在沪深股指上进行的实验表明,相较于传统的模型,文中提出的方法可以将股市较大拐点预测的准确度提高约10％,并在6个月的回测实验中准确率依旧保持在70％左右的水准,从而证明了模型的有效性.     

基于Agent的投资者情绪对于股市演化行为仿真研究

基于双向拍卖机制作为价格生成机制,应用遗传算法来进化预测规则,建立了中国股市的人工金融市场模型,并在此基础上研究了投资者情绪对于市场演化行为的影响。研究结果表明人工市场能够产生真实市场演化过程中的混沌动力学行为,并且市场演化行为随着投资者情绪的变化而变动。这一研究对挖掘中国股票市场的演化规律具有重要意义。     

多源异构数据融合驱动的股票指数预测研究

现代信息技术的广泛应用使得资本市场投资者能够获得更及时、更有价值的信息,也更容易受到金融论坛、专业投资网站的影响。融合资本市场的多源异构数据对股票指数进行预测成为该领域的研究热点。提出了一种基于多源异构数据的长短期神经网络（Long Short-Term Memory,LSTM）模型,通过对融合资本市场交易数据、技术指标数据、投资者情绪三种源数据的量化来预测股票指数的走势。提出了一种可以提取深度情感特征的卷积神经网络（Convolutional Neural Networks,CNN）情感分析模型,构建了投资者情绪特征模型。利用“上证50指数”数据进行实验,结果显示：LSTM模型的预测准确率比传统模型更为优秀,数据源的增加也对模型准确率的提升有较大贡献,验证了该方法的可行性和有效性。     

基于FinBERT-CNN的股吧评论情感分析方法

我国股市波动受投资者情绪变化影响较大,通过对股吧等金融交流平台上投资者的评论进行情感分析,能够帮助投资者更好地了解股票市场的变化.现有的情感分析方法是利用模型对股票评论集进行分析,但缺少优质的股票评论标注数据集用于模型训练,且单一模型提取股票评论特征较为片面,模型的准确性有待提高.该文针对股吧平台上的评论数据,提出一种...     

基于经验模态分解与投资者情绪的长短期记忆网络股票价格涨跌预测模型

针对传统的基于统计学的回归股票预测模型难以表征多个变量之间的关系,预测出的股票价格趋势误差较大,提出一种基于经验模态分解（EMD）与投资者情绪的长短期记忆（LSTM）神经网络股票价格涨跌预测模型。首先,将股票收盘价通过EMD分解得到若干个具有不同时间尺度的局部特征信号的本征模函数（IMF）;其次,通过引入改进的股票领域情感词典,对东方财富网股吧的帖子,进行上一个股票交易日收盘后和下一个股票交易日开盘前的投资者情感分析,得到下一个股票交易日的投资者情绪指标;最后,将基础的股票基本行情数据、经过EMD得到的IMF以及投资者情绪指标加入LSTM神经网络预测下一个交易日的股票涨跌。仿真实验结果表明,在2019年1月至2021年9月的牧原股份（002714）股票数据上,与单独使用LSTM模型相比,改进后的LSTM模型的预测准确率提高了12.25个百分点,在预测为涨的F1值和预测为跌的F1值上分别提高了1.2个百分点和25.21个百分点。由此可见,基于EMD与投资者情绪的LSTM股票价格涨跌预测模型有效提高了预测精度,为股票市场的涨跌预测提供了一种有效的实验方法。     

基于经验模态分解与投资者情绪的Attention-BiLSTM股价趋势预测模型

股票价格的变动是投资者在股票市场关注的焦点，所以股价趋势预测一直是量化投资研究的热门话题。传统的机器学习预测模型难以处理非线性、高频率、高噪声的股价时间序列，使得股票价格趋势的预测精度低。为了提高预测精度，针对股票价格数据的时序性特征，提出用结合经验模态分解（EMD）、投资者情绪和注意力机制的双向长短期记忆神经网络来对股票价格进行涨跌预测。首先使用经验模态分解算法提取股票价格时间序列在不同时间尺度上的特征，并通过构建金融情感词典来提取上一个股票交易日收盘后至下一个交易日开盘前文本的投资者情绪指标，最后使用注意力机制优化的BiLSTM模型对下一个股票交易日进行涨跌预测。在股票价格序列的数据集上进行实验，结果表明，改进后的BiLSTM模型较改进前的BiLSTM模型，准确率从58.50%提升至71.26%；预测为涨的精确率从58.20%提升至70.06%，预测为跌的精确率从59.34%提升至72.36%；预测为涨的召回率从59.85%提升至73.41%，预测为跌的召回率从57.73%提升至69.11%；预测为涨的F1值从58.60%提升至71.61%，预测为跌的F1值从58.08%提升至70...     

基于支持向量机的股价反转点预测

股票市场瞬息万变,股价反转点对投资者进行投资决策起着至关重要的作用。技术分析能够揭示股价反转的某些特征,但是使用单一技术指标预测股价反转点的召回率和准确率不高。本文提出了一种使用支持向量机（SVM）对技术指标组合进行数据挖掘,从而实现预测股价反转点的方法,实验结果表明该方法较使用单一技术指标进行反转点预测在召回率和准确率方面都有极大的提高。     

基于多重注意力机制的图神经网络股市波动预测方法

股票市场是金融市场关键组成部分,因此对股票市场波动的研究对合理化控制金融市场风险、提高投资收益提供了重要支持,一直以来都是学术界和相关业界的关注焦点,然而,股票市场会受到各种因素的影响。面对股票市场中多源化、异构化的信息,如何高效挖掘、融合股票市场的多源异构数据具有挑战性。为了充分解释不同信息及信息间相互作用对于股票市场价格波动的影响,提出一种基于多重注意力机制的图神经网络来预测股票市场的价格波动。首先,引入关系维度构建股票市场交易数据和新闻文本的异构子图,并利用多重注意力机制实现图数据的融合;其次,通过图神经网络门控循环单元（GRU）进行图分类,在此基础上完成对股票市场中上证综合指数、沪深300指数、深证成份指数这三个重要指数波动的预测。实验结果表明,从异构信息特性角度,相较于股票市场交易数据,股市新闻信息对于股票价格影响存在滞后性;从异构信息融合角度,所提方法与支持向量机（SVM）、随机森林、多核k-means （MKKM）聚类等算法相比,预测准确率分别提升了17.88个百分点、30.00个百分点和38.00个百分点,并进行了模型交易策略的量化投资模拟。     

结合公司财务报表数据的股票指数预测方法

股票市场参与者的所有市场活动综合影响着股票市场的变化,使股票市场的波动充满复杂性,也使得准确预测股票价格成为难题。在这些影响股市变化的活动中,财务披露是预测股票指数变化的一种吸引人的且具有潜在财务回报的手段。为了应对股票市场的复杂变化,提出一种结合公司披露的财务报表数据进行股票指数预测的方法。该方法首先对股票指数历史数据和公司财务报表数据进行预处理,主要是对公司财务报表数据生成的高维矩阵进行降维,然后用双通道的长短期记忆（LSTM）网络对归一化后的数据进行预测研究。在上证50指数和沪深300指数数据集上的实验结果表明,该方法的预测效果优于仅使用股票指数历史数据的预测效果。     

基于微博的股票投资者未来情感倾向识别研究

近年来,微博越来越受到网络用户的青睐,成千上万的用户通过发布微博共享他们的观点和情感。其中,有大量带有情感倾向(认为某事物“好”或“坏”)的微博,这些微博反映了作者的情绪。投资者情绪(investor sentiment)是研究经济市场走向的重要指标,行为金融学认为股票投资者情绪影响投资者决策,进而影响股票市场,而反映股票投资者情绪的重要指标是投资者对股票市场未来行情的情感倾向(认为股票市场未来行情“好”或“坏”)。通过对新浪微博(目前最大的中文微博平台)上股票投资者发布的文本进行情感信息方面的分析与研究,提出了一种自动识别股票投资者未来情感倾向的方法。该方法分为两级识别,第一级是:识别出微博中包含未来情感的句子;第二级是:将第一级识别出来的包含未来情感的句子分为正面评论(看涨)和负面评论(看跌)。实验结果表明,所提方法对自动识别股票投资者的未来情感倾向达到了非常好的效果。     

面向股票新闻的情感分类方法

互联网新闻资讯对证券市场和投资者有举足轻重的影响,新闻进行情感分类后再展示给用户,可以帮助投资者迅速做出投资决定.从文本分类的基本方法出发,实现了基于N-gram 统计模型的新词发现方法,并将所得结果用于构建中文分词词典和情感词典.同时引入评价理论,并用朴素贝叶斯、K 近邻和支持向量机3 种方法进行股票新闻标题的情感分类实验.所用实验数据来自2009 年“新浪财经”共计23 万余条的新闻标题,结果表明二分类的准确率最高可达82.9%.此外,还实现了一个原型系统用于展示股票新闻的分类结果.     

基于遗传算法的股票市场选择模型

为提高投资者在股票市场的收益,解决在证券投资中股票选择这一重要问题,提出一种基于遗传算法的股票选择模型。算法以上市公司的财务指标为样本特征,为克服K-means算法的不稳定性,采用基于遗传算法的K-means算法对同一板块股票进行聚类分析,剔除财务指标较差的一类中的股票。对筛选条件编码,为解决传统遗传算法处理复杂问题时存在的过早收敛现象,提出改进的遗传算子,利用改进的遗传算法寻找使股票市场投资收益最大化的选股模型参数。实验结果表明,该算法在股票选择上具有较好的效果,可供市场投资者借鉴。     

A double-threshold GARCH model of stock market and currency shocks on stock returns

International integration of financial markets provides a channel for currency movements to affect stock prices. This paper applies a four-regime double-threshold GARCH (DTGARCH) model of stock market returns to investigate empirically the effects of daily currency movements on five stock market returns, namely in Taiwan, Singapore, South Korea, Japan and the USA. The asymmetric reactions of the mean and volatility stock returns in five markets to stock market and foreign exchange news are investigated using linear and nonlinear models. We discuss a four-regime DTGARCH model, which allows for asymmetry in both the conditional mean and conditional variance simultaneously by using two threshold variables to analyze stock market reactions to different types of information (that is, positive and negative news) that are generated from stock and foreign exchange markets. By applying the four-regime DTGARCH model, this paper finds that the interactions between the information of stock and foreign exchange markets lead to asymmetric reactions of stock returns and their associated variability. The empirical results show that international fund managers who invest in newly emerging stock markets need to evaluate the value and stability of domestic currencies as part of their stock market investment decisions.     

关联规则在股票板块联动分析中的应用

Apriori算法是关联规则挖掘中的经典算法,针对Apriori算法的不足进行了一些改进。新算法使用垂直数据格式,并改进了产生候选项的连接方法。为了研究股票板块的联动关系,将改进算法应用于股票板块指数分析中。实验结果表明,改进算法能快速发现板块之间的联动关系,对股市分析和投资决策有一定的指导作用。     

Application of the consensus method in a multiagent financial decision support system

Financial decisions are supported by different methods, based mainly on statistics, mathematics, behaviorism, artificial intelligence or experts’ opinion sentiment analysis. The main problem is that under conditions of risk and uncertainty predicting financial markets can be very difficult. This paper presents an approach to investment strategy design for a multiagent system which supports investment decisions on the stock market. Individual components and functionalities of the multiagent financial decision support system method have been briefly described. On the basis of decisions generated by agents, the Supervisor Agent uses a consensus method to generate a satisfactory rate of return and reduce the level of risk associated with investing in a financial instrument. Verification of the effectiveness of the strategy has been conducted using investments on the Warsaw Stock Exchange.     

AN ADAPTIVE EXPECTATION GENETIC ALGORITHM BASED ON ANFIS AND MULTINATIONAL STOCK MARKET VOLATILITY CAUSALITY FOR TAIEX FORECASTING

The stock market is a highly complex and dynamic system, and forecasting stock is complicated and difficult. Successful prediction of stock prices may promise attractive benefits; therefore, stock market forecasting is important and of great interest. The economy of Taiwan relies on international trade deeply and the fluctuations of international stock markets impact Taiwan's stock market to certain degree. It is practical to use the fluctuations of other stock markets as forecasting factors for forecasting on the Taiwan stock market. Further, stock market investors usually make short-term decisions based on recent price fluctuations, but most time series models use only the last period of stock price in forecasting. In this article, the proposed model uses the fluctuations of other national stock markets as forecasting factors and employs an expectation equation method whose parameters are optimized by a genetic algorithm (GA) joined with an adaptive network–based fuzzy inference system (ANFIS) model to forecast the Taiwan stock index. To evaluate the forecasting performance, the proposed model is compared with Chen's model and Yu's model. The experimental results indicate that the proposed model is superior to the listing methods (Chen's model and Yu's model) in terms of root mean squared error (RMSE).     

Stock prediction:an event-driven approach based on bursty keywords

There are many real applications existing where the decision making process depends on a model that is built by collecting information from different data sources. Let us take the stock market as an example. The decision making process depends on a model which that is influenced by factors such as stock prices, exchange volumes, market indices (e.g. Dow Jones Index), news articles, and government announcements (e.g., the increase of stamp duty). Yet Nevertheless, modeling the stock market is a challenging task because (1) the process related to market states (rise state/drop state) is a stochastic process, which is hard to capture using the deterministic approach, and (2) the market state is invisible but will be influenced by the visible market information, like stock prices and news articles. In this paper, we propose an approach to model the stock market process by using a Non-homogeneous Hidden Markov Model (NHMM). It takes both stock prices and news articles into consideration when it is being computed. A unique feature of our approach is event driven. We identify associated events for a specific stock using a set of bursty features (keywords), which has a significant impact on the stock price changes when building the NHMM. We apply the model to predict the trend of future stock prices and the encouraging results indicate our proposed approach is practically sound and highly effective.     

Stock prediction: an event-driven approach based on bursty keywords

There are many real applications existing where the decision making process depends on a model that is built by collecting information from different data sources. Let us take the stock market as an example. The decision making process depends on a model which that is influenced by factors such as stock prices, exchange volumes, market indices (e.g. Dow Jones Index), news articles, and government announcements (e.g., the increase of stamp duty). Yet Nevertheless, modeling the stock market is a challenging task because (1) the process related to market states (rise state/drop state) is a stochastic process, which is hard to capture using the deterministic approach, and (2) the market state is invisible but will be influenced by the visible market information, like stock prices and news articles. In this paper, we propose an approach to model the stock market process by using a Non-homogeneous Hidden Markov Model (NHMM). It takes both stock prices and news articles into consideration when it is being computed. A unique feature of our approach is event driven. We identify associated events for a specific stock using a set of bursty features (keywords), which has a significant impact on the stock price changes when building the NHMM. We apply the model to predict the trend of future stock prices and the encouraging results indicate our proposed approach is practically sound and highly effective.