Affiliations: [a] Department of Computer Science and Information Engineering, Tamkang University, Taipei, Taiwan | [b] Institute of Manufacturing Information & Systems, Department of Computer Science & Information Engineering, National Cheng Kung University Tainan, Taiwan | [c] Department of Computer Science and Intelligent Systems, Osaka Prefecture University, Osaka, Japan
Correspondence:
[*]
Corresponding author: Chao-Chun Chen, Institute of Manufacturing Information & Systems, Department of Computer Science & Information Engineering, National Cheng Kung University Tainan, Taiwan. E-mail:[email protected]
Abstract: Portfolio optimization is always an attractive topic for research. In our
previous approach, we proposed a method for mining a group stock portfolio
that used grouping genetic algorithms. The derived group stock portfolio
represents stocks in the same group that may have similar properties;
consequently, a variety of stock portfolios could be offered to investors.
However, the evaluation process used by this previous approach is time
consuming when the number of stocks or groups increases. To address this
problem, the map-reduce technique is considered. Map-reduce is a well-known
approach for speeding up the mining process. This paper proposes a
map-reduce-based approach to mine groups of stock portfolios and speed up
the evolution process while still achieving results as similar as possible
to the previous approach. Here, a chromosome represents a mapper number, a
group number, a stock part and a portfolio part. Utilizing the mapper
number, the chromosomes in a population are divided into subsets and sent to
respective mappers, while the reducers execute fitness evaluation and
genetic operations. The evolution process is repeated until the terminal
conditions are reached. Finally, experiments on a real dataset are conducted
to demonstrate the efficiency of the proposed approach.
