What is the meaning of "unexplained" in unexplained variance or residual sum of squares? I understand the formula of RSS and RSE but it confuse me every time I read unexplained variance. I don't understand why the term unexplained is use. 
 A: In the case of a linear model we use a linear combination of covariates to try and describe a response variable (i.e. we regress the covariate onto the response).
By definition, there is variation in the response. The covariates have some correlation (hopefully) with the response variable, which is meant to encapsulate something of the relationship between the two. 
This relationship/correlaion can be thought of as the variation in the response that we can explain (i.e. a 1 unit change in our covariate $x$$i$ will correspond to some quantified change in our response $y$). 
The amount of variation that exists in our response variable is the total sum of squares (TSS). Our model, ideally, explains some of this; the explained sum of squares (ESS) . 
However, typically our models do not explain all the variation that exists in our response variable - there is some theoretically random variation left over that our covariates can't explain; the residual sum of squares (RSS).  Hence there is some unexplained variance.
Please feel free to ask if anything above is unclear, and hopefully this answers the question you were asking!
