Abstract
“Robust Parameter Design（RPD）” is used to design parameters to construct robust functions under different noises. On the other hand, in deciding on an investment portfolio, which is a combination of investment assets of an individual or a company, efficient asset allocation is important; increase profitability as well as prevent a change of profitability. On the basis of a preceding study that used RPD for stock portfolio selection, this study improves the technique of that study and proposes a more efficient technique to select such portfolio. Thus it shows a further practical research of RPD. A stock investment simulation based on the proposed technique enabled selection of a much more efficient stock portfolio than those selected using the existing technique. It can be considered that the most effective improvement is to use the Signal-to-Noise ratio of dynamic characteristics as a performance index in order to decide optimal levels.
