Comments - Multi-Regression in R (Exxon Mobil stock price ~ WTI, Gas, and S&P500) - Data Science Central2019-09-20T01:09:25Zhttps://www.datasciencecentral.com/profiles/comment/feed?attachedTo=6448529%3ABlogPost%3A427270&%3Bxn_auth=noGregory:
Very good tutorial,â€¦tag:www.datasciencecentral.com,2016-05-30:6448529:Comment:4306652016-05-30T16:25:57.408ZAlberto Gaidyshttps://www.datasciencecentral.com/profile/AlbertoGaidys
Gregory:<br />
<br />
Very good tutorial, thank you. There is a 4th more and more likely possibility for the negative coefficient for the price of Natural Gas: its strong correlation with the price of WTI Oil, causing what in Multiple Regression is called as "Multicolinearity". It is well documented in the literature that the presence of Multicolinearity among the independent variables causes their coefficients to loose precision individually. The regression is still valid, but one can't infer directâ€¦
Gregory:<br />
<br />
Very good tutorial, thank you. There is a 4th more and more likely possibility for the negative coefficient for the price of Natural Gas: its strong correlation with the price of WTI Oil, causing what in Multiple Regression is called as "Multicolinearity". It is well documented in the literature that the presence of Multicolinearity among the independent variables causes their coefficients to loose precision individually. The regression is still valid, but one can't infer direct relationships between the dependent and each independent variable from its coefficients anymore, only their joint effect. One possible solution, if you still wish to isolate the individual effect of each independent variable, would be to extract the 3 principal components from the multivariate sample using princomp, run a regression exactly as you did and roll back the coefficients from the transformation matrix -- this is the so called "Principal Component Regression". Or use 3 Independent Components for the regression through fastICA. Or yet you can use Ridge Regression or Partial Least Squares Regression, that are known to mitigate this effect. I hope this helps.<br />
<br />
A.Gaidys