%0 Journal Article
%T 基于粒子群优化算法的股票指数追踪及投资分析
Stock Index Tracking and Investment Analysis Based on Particle Swarm Optimization Algorithm
%A 陈鑫
%J Operations Research and Fuzziology
%P 3299-3313
%@ 2163-1530
%D 2023
%I Hans Publishing
%R 10.12677/ORF.2023.134333
%X 本文首先将二进制粒子群特征选择算法结合岭估计、Lasso估计和最小二乘估计利用成分股构建了三个关于上证50指数的指数追踪模型,三个模型都取得了非常优秀的追踪效果,其中无论是从追踪效果的角度还是投资角度出发,基于二进制粒子群特征选择算法结合最小二乘估计所得到的指数追踪模型都是最佳模型。然后利用三个指数追踪模型提取的成分股计算对上证50趋势的影响排名,其中贵州茅台的影响排名最高。最后在一定条件下进行投资行为的模拟和收益分析,发现基于三个指数追踪模型所提取的优质成分股进行投资可以获得比直接投资上证50指数更高的收益,达到了降低投资成本和投资风险的目的,为证50指数的投资提供科学合理的建议。
In this paper, we first combine the binary particle swarm feature selection algorithm with ridge estimation, Lasso estimation and least square estimation to construct three index tracking models on SSE 50 Index using constituent stocks. The three models have achieved excellent tracking effects. The exponential tracking model based on binary particle swarm optimization feature selection algorithm and least squares estimation is the optimal model, whether from the perspective of tracking effect or investment. Then we use the constituent stocks extracted from the three index tracking models to calculate the impact ranking on the Shanghai Stock Exchange 50 trend, of which Kweichow Moutai ranks the highest, Finally, under certain conditions, the simulation and return analysis of investment behavior are carried out, and it is found that the investment of high-quality constituent stocks extracted based on the three index tracking models can obtain higher returns than direct investment in the SSE 50 Index, achieving the purpose of reducing investment costs and investment risks, and providing scientific and reasonable suggestions for the investment of the 50 Index.
%K 指数追踪,二进制粒子群,特征选择,岭估计,Lasso估计,最小二乘估计
Exponential Tracking
%K Binary Particle Swarm
%K Feature Selection
%K Ridge Estimation
%K Lasso Estimation
%K Least Squares Estimation
%U http://www.hanspub.org/journal/PaperInformation.aspx?PaperID=70340