Exclusion of risky values of trading models parameters in investment decisions

Anticipated profits and the associated expected risks are the two main criteria governing appropriate investment decisions. In case of daily trading decisions in financial markets, investor ought to accept risk in order to make profit from trading. Several trading models have been introduced each of these need the appropriate selection of its parameter's value. Present work is concerned with excluding risky values of these parameters. One of the well-known models in managing investment risk is Kelly Criterion. This criterion enables investors to decide rationally the proportion of their investment that could be released for trading. Furthermore, the current paper focuses on the capability of Kelly criterion to differentiate between risky and safe investments. The most contributing spectral analysis components are used for effectively predicting Kelly values. The effectiveness of the proposed method for different trading models is demonstrated by implementation on actual trading data. The proposed method shows a considerable decrease in Mean Absolute Error (MAE) comparing with moving average method.