Paper Info
Title
A minimax model of portfolio optimization using data mining to predict interval return rate
Abstract
In 1950s, Markowitzs first proposed portfolio theory based on a mean-variance (MV) model to balance the risk and profit of decentralized investment. The two main inputs of MV are expected return rate and the variance of expected return rate. The expected return rate is an estimated value which is often decided by experts. Various uncertainty of stock price brings difficulties to predict return rate even for experts. MV model has its tendency to maximize the influence of errors in the input assumptions. Some scholars used fuzzy intervals to describe the return rate. However, there were still some variables decided by experts. This paper proposes a classification method to find the latent relationship between the interval return rate and the trading data of a stock and predict the interval of return rate without consulting any expert. Then this paper constructs the portfolio model based on minimax rule with interval numbers. The evaluation results show that the proposed method is reliable.
Year
DOI
Venue
2014
10.1109/FUZZ-IEEE.2014.6891693
FUZZ-IEEE
Keywords
Field
DocType
mean-variance model,fuzzy set theory,interval number,classification,mv model,fuzzy intervals,classification method,pattern classification,portfolio,financial data processing,interval return rate prediction,stock price uncertainty,portfolio theory,share prices,decentralized investment,stock markets,minimax techniques,data mining,minimax,portfolio optimization,minimax model,security,data models,predictive models,linear programming,investment
Econometrics,Minimax,Actuarial science,Computer science,Portfolio,Artificial intelligence,Rate of return,Modern portfolio theory,Efficient frontier,Portfolio optimization,Rate of return on a portfolio,Machine learning,Expected return
Conference
ISSN
Citations 
PageRank 
1544-5615
0
0.34
References 
Authors
0
2
Name
Order
Citations
PageRank
Meng Yuan100.68
Junzo Watada241184.53