Advances and Applications in Statistics
Volume 46, Issue 1, Pages 57 - 78
(July 2015) http://dx.doi.org/10.17654/ADASJul2015_057_078 |
|
APPLICATION OF BAYESIAN INFERENCE FOR THE ANALYSIS OF STOCK PRICES
Juliet Gratia D’Cunha, K. Aruna Rao and T. Mallikarjunappa
|
Abstract: In the analysis of stock market data, it is important to predict the future price of various scripts. Generally, stock prices are analyzed using log transformation. This leads to the assumption that distribution of stock price is lognormal. In this paper we develop Bayes estimator and credible intervals for the mean of the lognormal distribution. The advantage of Bayesian inference is that future stock prices can be predicted using past information as well as the belief of the investor. The validity of proposed method is examined by analyzing the average daily prices (as reported by National Stock Exchange of India limited (NSE)) of 18 scripts listed in NSE for the period of 3 months from October 1st to December 31st, 2013. These scripts belong to 9 sectors namely automobile, bank, cement, finance, FMGC, IT software, pharmaceuticals, real estate, and telecommunication services. Out of these 18 scripts, 11 scripts belong to the list of Bombay Stock Exchange (BSE) sensex companies. The validity of the procedure is examined using training and validation data sets. The conclusion is that the data of previous 5 days is sufficient to accurately predict the stock price of the 6th day. A rule has also been suggested for making use of the forecasted value for taking decision by the market analyst and the stock broker and can also be incorporated in automated forecasting. |
Keywords and phrases: Bayesian inference, stock prices, lognormal distribution, credible intervals, right invariant prior, left invariant prior. |
|
Number of Downloads: 398 | Number of Views: 1341 |
|