摘要
多因子选股模型作为一种量化投资选股策略,可避免交易者个人主观意念的干扰。券商提供的股票公司研报往往包含着丰富及时的信息,通过获取研报信息并从中提取相关特征指标,构造基于研报的量化选股模型,能为投资者提供更多的参考。本文首先从收集到的股票研报中初步提取出29个特征指标,进行pearson相关性分析,得到相关性较弱的10支股票。接着利用多元线性回归与等权重法,对这些股票的特征指标进行打分,提取出最重要的10个特征指标作为最终有效因子。而后10支股票分别列出多元线性回归方程,从而绘制出曲线图来分析股票净利率走势,筛选出高利润和一般利润股票。接着在对应多因子模型的基础上,以净利润成长性最高为目标,根据相关特征指标的发展趋势对一般收益股票进行分组。通过以上分析给出对这10支股票的下年度的持仓策略。
As a quantitative investment stock selection strategy, multi factor stock selection model can avoid the interference of traders’ personal subjective ideas. The Research Report of stock companies provided by securities companies often contains rich and timely information. By obtaining the research report information and extracting relevant characteristic indicators, constructing a quantitative stock selection model based on the research report can provide more references for investors. Firstly, this paper preliminarily extracts 29 characteristic indexes from the collected stock research reports, carries out Pearson correlation analysis, and obtains 10 stocks with weak correlation. Then, the characteristic indexes of these stocks are scored by using multiple linear regression and equal weight method, and the most important10 characteristic indexes are extracted as the final effective factor. Then 10 stocks list multiple linear regression equations respectively, so as to draw a curve to analyze the trend of stock net interest rate and screen out high profit and general profit stocks. Then, based on the corresponding multi factor model, with the highest net profit growth as the goal, the general return stocks are grouped according to the development trend of relevant characteristic indicators. Through the above analysis, the position strategy of these 10 stocks in the next year is given.
作者
周嘉灏
张明福
冷鸿杰
Zhou Jiahao;Zhang Mingfu;Leng Hongjie(South China Agricultural University,College of Electronic Engineering(Artificial Intelligence Academy),Guangzhou 510642,China)
出处
《科学技术创新》
2022年第5期161-164,共4页
Scientific and Technological Innovation
关键词
多因子模型
量化选股
多元线性回归
Multi factor model
Quantitative stock selection
Multiple linear regression