Wavelet Transform Based Algorithm to Investigate Periodicities and Correlation of Stocks in Sensex and Various Sectors of Bombay Stock Exchange

Basak, Suman (2014) Wavelet Transform Based Algorithm to Investigate Periodicities and Correlation of Stocks in Sensex and Various Sectors of Bombay Stock Exchange. Masters thesis, Indian Institute of Science Education and Research Kolkata.

PDF (MS dissertation of Suman Basak) Suman_Basak.pdf - Submitted Version Restricted to Repository staff only Download (4MB)

Abstract

Wavelet Analysis may be perceived to have caught great interest of Mathematicians and Physicists mainly because they satisfy certain specific mathematical requirements. In this thesis I have used Wavelet transform method in mainly time series data. In the case of BSE Sensex the cross correlation matrix obtained showed few patches of both correlated and uncorrelated companies. However these patches were not localized in regions and were spreaded in randomly. This is due to the fact that BSE Sensex has companies from different sectors. Besides there are a mixture large cap,mid cap and small cap companies. So the correlation amongst them are not that high. However when we look at the correlation matrices obtained for the four sectors of BSE sensex. We find that the sector specific companies are very highly correlated with one another. Besides you can often see clear boundaries separating the four sectors. Also in case of sectorial analysis we often had high inter-sector correlation too. Now the companies in the various sectors involved were mostly large cap companies. This shows that large cap companies from different sectors also follows similar trends and thereby control the overall market behaiviour. But in case of BSE Sensex we had a mixture of large, medium and small cap companies and thereby correlation among them were low.

Actions (login required)

View Item