«上一篇/Previous Article|本期目录/Table of Contents|下一篇/Next Article»

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

33

期数:

2023年04期

页码:

190-197

栏目:

新型计算应用系统

出版日期:

2023-04-10

文章信息/Info

Title:

Research on Application of Dimension Reduction Method of High Dimensional Sequence Data in Securities Market

文章编号:

1673-629X(2023)04--0190-08

作者:

陈 赛¹ ; 刘文杰¹ ; 黄国耀¹ ; 卢凌峰¹ ; 李华康2 ; 孙国梓^1*

1. 南京邮电大学 计算机学院,江苏 南京 210023;

2. 西交利物浦大学 人工智能与高级计算学院(太仓),江苏 苏州 215123

Author(s):

CHEN Sai1 ; LIU Wen-jie1 ; HUANG Guo-yao1 ; LU Ling-feng1 ; LI Hua-kang2 ; SUN Guo-zi1*

1. School of Computer Science,Nanjing University of Posts and Telecommunications,Nanjing 210023,China;

2. School of Artificial Intelligence and Advanced Computing,XJTLU Entrepreneur College ( Taicang) ,Suzhou 215123,China

关键词:

数据挖掘; 易经; 特征筛选; 证券预测; 机器学习

Keywords:

data mining; IChing; feature screening; securities forecasting; machine learning

分类号:

TP391

DOI:

10. 3969 / j. issn. 1673-629X. 2023. 04. 028

摘要:

证券市场数据分析与预测,作为一个经典的大数据分析场景,很多数据挖掘方法已经在本领域得到实际应用。 但是鉴于企业本身情况的变化以及证券市场的人为操作等情况,现有的各种大数据挖掘方法无法应对不可见或者未出现的情况,为此论文探索使用易经方法,将其应用在证券市场的数据挖掘和分析预测。 利用数据挖掘进行特征筛选、数据降维,通过滑动时间窗、随机森林、三才映射等方法实现传统易经体系中的断卦步骤,将易经概念、规则抽象成算法并对卦辞分类,由解卦算法得出预测值。 与先前的预测模型相比,该模型融合易经预测体系与机器学习,充分利用了证券市场的场景特征与历史数据,最终对证券市场平稳、上升、下跌三种发展趋势进行预测。 使用 10 年内股票证券交易公共数据集进行实验,准确率优于 SVM、XGBoost 等流行的机器学习算法,并在分行业建模中进一步提升了效果。

Abstract:

As a classic big data analysis scenario,many data mining methods have been practically applied in the field of securities marketdata analysis and predicting. However,various existing data mining methods cannot deal with invisible or non-existent situations causedby the company diversification and the human intervention. Therefore, IChing method is used in data mining,analysis and prediction ofsecurities market. Data mining is used for feature selection and data dimension reduction,sliding time window,random forest,three-waymapping and other methods are adopted to realize the steps of hexagrams breaking in the traditional IChing system,and the concepts and rules of IChing are abstracted into an algorithm and the hexagrams are classified,and the predicted values are obtained from the hexagramsbreaking algorithm. Compared with the previous prediction model, the proposed model integrates the IChing prediction system andmachine learning,makes full use of the scene characteristics and historical data of the securities market,and finally predicts the stable,rising and falling development trends of the securities market. Using the public data set of stock exchange within 10 years for experiment,the accuracy is better than that of SVM,XGBoost and other popular machine learning algorithms,and the effect is further improved in thesub-industry modeling.

常用功能

更新日期/Last Update: 2023-04-10