Merge pull request #1234 from slds-lmu/fix-ex-eval3

fix typo

exercises/evaluation/evaluation_3.html

@@ -3538,7 +3538,7 @@ <h2 class="anchored" data-anchor-id="exercise-1-roc-metrics">Exercise 1: ROC met
 <summary>
 <strong>Solution</strong>
 </summary>
-<p>We need the table sorted by score (descending). Finding that <span class="math inline">\(\frac{1}{n_{+}} = \frac{1}{n_{+}} = 0.2\)</span>, we follow the algorithm described in the lecture slides:</p>
+<p>We need the table sorted by score (descending). Finding that <span class="math inline">\(\frac{1}{n_{+}} = \frac{1}{n_{-}} = 0.2\)</span>, we follow the algorithm described in the lecture slides:</p>
 <ol type="1">
 <li><span class="math inline">\(c = 1\)</span> <span class="math inline">\(~ \Longrightarrow ~\)</span> we start in <span class="math inline">\((0,0)\)</span> and predict everything as negative, so TPR 0 and FPR 0.</li>
 <li><span class="math inline">\(c = 0.625\)</span> <span class="math inline">\(~ \Longrightarrow ~\)</span> TPR <span class="math inline">\(0 + \frac{1}{n_{+}} = 0.2\)</span> and FPR 0 (obs 6 correctly classified).</li>

exercises/evaluation/evaluation_3.qmd

@@ -145,7 +145,7 @@ Draw the ROC curve and interpret it.
 <summary>**Solution**</summary>
 
 We need the table sorted by score (descending).
-Finding that $\frac{1}{\np} = \frac{1}{\np} = 0.2$, we follow the algorithm described in the lecture slides:
+Finding that $\frac{1}{\np} = \frac{1}{\nn} = 0.2$, we follow the algorithm described in the lecture slides:
 
  1. $c = 1$ $~ \Longrightarrow ~$ we start in $(0,0)$ and predict
   everything as negative, so TPR 0 and FPR 0.