simplify notation

exercises/evaluation/ex_rnw/ex_of_uf.Rnw

@@ -1,9 +1,8 @@
 Assume a polynomial regression model with a continuous target variable $y$ and 
 a continuous, $p$-dimensional feature vector $\xv$ and polynomials of degree 
 $d$, i.e., 
-$$\fxi = \sum_{j=1}^p \sum_{k=0}^d \theta_{j,k} (\xi_j)^k,$$ and 
-$\yi = \fxi + \epsi$ where the $\epsi$ are iid with $\var(\epsi)=\sigma^2 \ 
-\forall i\in \{1, \dots, n\}$. 
+$$\fxi = \sum_{j=1}^p \sum_{k=0}^d \theta_{j,k} \left(\xi_j\right)^k,$$ 
+$\forall i\in \{1, \dots, n\}$. 
 
 \begin{enumerate}[a)]
   \item For each of the following situations, indicate whether we would 
@@ -17,7 +16,7 @@ $\yi = \fxi + \epsi$ where the $\epsi$ are iid with $\var(\epsi)=\sigma^2 \
     observations $n$ is small.
     \item[(iii)] The true relationship between the features and the response is 
     highly non-linear.
-    \item[(iv)] The variance of the error terms, $\sigma^2$, is extremely high.
+    \item[(iv)] The data are very noisy.
   \end{enumerate}
   \item Are overfitting and underfitting properties of a learner or of a
   fixed model? Explain your answer.