计算机与现代化

• 数据挖掘 • 上一篇    下一篇

基于k-means聚类的股票KDJ类指标综合分析方法

  

  1. (中央财经大学信息学院,北京102206)
  • 收稿日期:2018-04-17 出版日期:2018-10-26 发布日期:2018-10-26
  • 作者简介:李娜(1992-),女,山东博兴人,中央财经大学信息学院硕士研究生,研究方向:数据挖掘,金融大数据; 毛国君(1966-),〖JP2〗男,教授,博士,研究方向:数据挖掘,分布式计算; 邓康立(1994-),男,硕士研究生,研究方向:数据挖掘,金融大数据。
  • 基金资助:
    国家自然科学基金资助项目(61773415)

K-means-based KDJ Integrated Analyzing Methods for Stock Transactions

  1. (School of Information, Central University of Finance and Economics, Beijing 102206, China)
  • Received:2018-04-17 Online:2018-10-26 Published:2018-10-26

摘要: 股票技术分析是证券分析的常用手段之一,目前的股票技术分析主要存在2个问题:1)都是从某个角度进行单维度分析,投资决策有较大偏差;2)任何单一的技术指标都有其局限性,需要相互补充才能更好进行投资决策。针对这些问题,本文讨论如何利用数据挖掘技术进行股票多维度综合分析问题。首先,分析数据挖掘应用到股票分析中可以解决的问题及可能面临的挑战;其次,提出一种基于数据挖掘聚类方法的选股模型;最后,对1364只上证股票进行实证分析,形成对股票的随机指标K、D、J等的综合挖掘结果。

关键词: 数据挖掘, 聚类分析, 股票技术分析, 随机指标, k-均值算法

Abstract: Stock technical analysis is one of the means of securities analysis. There are two main problems in the current stock technical analysis. Firstly, one technical index is always analyzed in a dimension, and so the general investors are difficult to put them together to form an investment decision; secondly, any single technical index has its limitations, and so they need been integrated to make better investment decisions. In response to these major issues, this article discusses how to use the data mining technology for multi-dimensional comprehensive analysis of stocks. First of all, it analyzes the problems that data mining can solve in stock analysis and its possible challenges. Secondly, a stock selection model based on data mining clustering methods is proposed. Finally, using the 1364 Shanghai Stocks, some empirically analyzing results are given.

Key words: data mining, clustering analysis, technical analysis of stock, KDJ index, k-means algorithm

中图分类号: