An investor would like to build a balanced portfolio with stocks representing different sectors. Several researchers have attempted the portfolio selection problem by different methods. Many of these methods consider companies of different sectors together. However, it can be argued that the attributes affecting the company's growth vary for different sectors. Therefore, it is advisable to compare a company with the companies of the same sector. There are many options for the selection of a stock from a particular sector. A stock ranking method is proposed by using MADM methods based on overall performance under a stochastic environment. Of many MADM methods, SAW, AHP, TOPSIS, and VIKOR are applied. Usually, Euclidean distances (2-norm) are considered in the implementation of TOPSIS and VIKOR methods. In this work, this norm is generalized to p-norm, where p > 1. The model is tested for 13 companies in the field of Information Technology sector (IT) listed on National Stock Exchange in India and 13 criteria as performance indicators of a company. A MATLAB GUI system is developed and the results are obtained for several values of p in case of TOPSIS and VIKOR methods besides other methods. As the result indicates, the ordering is not much affected by different values of p in certain range. Moreover, higher values of p have adverse effect on the ordering. The proposed model is able to provide better information on the overall performance of a particular stock in comparison with its peers. The results obtained by various methods clearly separate good companies from inferior companies though the exact ordering slightly differs.

A predictability index for time series of a financial market vector consisting of chosen market parameters is suggested providing a measure of long range predictability of the market. It is based on fractional Brownian motion that includes Brownian motion as a particular case followed by the time series of financial market parameters. By analyzing respective time series, these indices are computed for parameters like volatility, FII investments in the local market, IIP numbers, CPI numbers, Dow Jones Index, different stock market indices, currency rates, and gold prices.

A predictability index for time series of a financial market vector consisting of chosen market parameters is suggested providing a measure of long range predictability of the market. It is based on fractional Brownian motion that includes Brownian motion as a particular case followed by the time series of financial market parameters. By analyzing respective time series, these indices are computed for parameters like volatility, FII investments in the local market, IIP numbers, CPI numbers, Dow Jones Index, different stock market indices, currency rates, and gold prices.