the sample size (n=6) is very small to run regression model.

In your case, the data is not enough.

You mean that you have, say, six subjects, and you measured 8 different things about each? What are you trying to determine?

have a look at PLS (Partial Least Squares), but I agree with Amin that it is not a large sample you have.

Best.

Hello Naresh,.

You'll need at least 2 more cases than predictors in order to estimate a non-trivial solution. If you have just 1 more case than predictors, you'll get an artificially perfect model-data fit (that is exceedingly unlikely to generalize), but no tests of significance.

All the others are correct in suggesting this is far too few cases for a meaningful analysis. One guideline you could consider would be a minimum of 10-20 cases per predictor (independent variable). If you want a more precise estimate of suitable sample size for your research goals, try using a program like G*Power (free for windows and mac OS versions: http://www.gpower.hhu.de/).

Good luck with your work!

I take a chance to thank you all for your fact-full information.

i take your valuable suggestion to increase my samples size.