基于KNN和神经网络算法的数据挖掘与预测模型研究  

Research on Data Mining and Prediction Model Basedon KNN and Neural Network Algorithm

作　　者：张迪[1] ZHANG Di(Anhui Vocational and Technical College of Industrial Economics,Hefei 230051,China)

机构地区：[1]安徽工业经济职业技术学院,安徽合肥230051

出　　处：《太原师范学院学报（自然科学版）》2023年第2期29-34,共6页Journal of Taiyuan Normal University:Natural Science Edition

基　　金：安徽省高校质量工程项目(2021jxtd047,2020sjjd020)。

摘　　要：数据挖掘是运用统计学、机器学习和数据库理论等多方面学科知识来对数据进行提炼获取的多学科领域.为了从大量的数据中提炼出有效的信息,拓宽数据挖掘技术在现实生活中的应用范围.利用数据挖掘中的最近邻法、支持向量机法、卷积神经网络法和CNN-LSTM混合神经网络法分别对股价数据进行分析建模和预测,并对所建立模型进行检验.结果表明,运用KNN法对股票价格的涨跌趋势进行短期预测是可行的,而在股票价格实时预测方面,按算法预测的准确度由小到大排序为:SVM法<CNN法<CNN-LSTM法.Data mining is a multidisciplinary field that uses statistics,machine learning and database theory to extract and obtain data.In order to extract effective information from a large amount of data,the application scope of data mining technology in real life is broadened.Using the nearest neighbor method,support vector machine method,convolutional neural network method and CNN-LSTM hybrid neural network method in data mining to analyze,model and predict the stock price data,and test the established model.The results show that it is feasible to use the KNN method to make short-term predictions on the rising and falling trends of stock prices.In terms of real-time forecasting of stock prices,the accuracy of the algorithm prediction is sorted from small to large:SVM method<CNN method<CNN-LSTM Law.

关 键 词：数据挖掘 股票价格 实时价格 涨跌趋势 预测 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程] F22[自动化与计算机技术—控制科学与工程]