We are working on developing a time series clustering algorithm for share market to group stocks based on similarity. One potential application is portfolio management. Can you please help me find more application of the same ?
Also volatility forecasting, or any forecasting in general. If you find that one stock has dynamics similar to for example 5 other stocks, you can use information from those other stocks to improve forecasts, whether volatility forecasts, or any other forecasts.
One way to investigate efficiency of portfolio diversification is to assess evidence of multivariate and bivariate nonlinear long-term relationship between the stock prices forming the portfolio. Nonlinear cointegration between stock prices imply portfolios in these markets are inefficient (systematic risk cannot be diversified away), as movement in the price of one market influence the movement in another market in a predictable direction and disproportionately. A good paper on this: "Analysis of portfolio diversification efficiency in emerging African stock markets" published in International J of Finance and Economics, issues 10, 2010.
Goodluck.
- Badges
- Science topic
- Similar topics
- Condensed Matter Physics
- Zero