I have a linear equation: as Y = -4 + 5X1 + 0.9X2 +0.5X3; Suppose correlation between X1 and X2 is 0.7. Since X1 has more predictive power than X2, regression picked X1 with more weightage. X2 was highly correlated with X1 and majority of the correlation of X2 with Y was expained by X1 in itself and hence X2 gets weightage only for the portion that was not explained by X1.
Now in order to read the equation, if I want to increase X2 by one unit, then equation says that Y increase by 0.9 considering everything else to be constant. But this cannot be the case as some portion in X2 was explained by X1 and therefore estimate of X2 does not capture it. I have a task in my hand to find the increment in value of Y2 when I increase X2 by a unit. How can I Achieve this?