Fixes from reviewer comments

articles/machine-learning/machine-learning-algorithm-cheat-sheet.md

@@ -71,7 +71,6 @@ Download and print the Machine Learning Algorithm Cheat Sheet in tabloid size to
 
 * Neural networks can be used with categorical variables by creating a **dummy variable** for each category and setting it to 1 in cases where the category applies, 0 where it doesn’t.
 
--->
 
 <!-- This is how you can embed a link in an image in HTML. Don't know how to do this in markdown.
 <a href="http://download.microsoft.com/download/A/6/1/A613E11E-8F9C-424A-B99D-65344785C288/microsoft-machine-learning-algorithm-cheat-sheet.pdf">

articles/machine-learning/machine-learning-algorithm-choice.md

@@ -322,7 +322,7 @@ And for those overachievers who wish to [specify their own network
 structure](http://go.microsoft.com/fwlink/?LinkId=402867), the
 possibilities are inexhaustible.
 
-## ![Boundaries learned by neural networks][6]
+![Boundaries learned by neural networks][6]
 ***The boundaries learned by neural networks can be complex and
 irregular***
 

articles/machine-learning/machine-learning-evaluate-model-performance.md

@@ -113,6 +113,7 @@ Figure 6. Binary Classification Confusion Matrix.
 Going back to the income classification problem, we would want to ask several evaluation questions that help us understand the performance of the classifier used. A very natural question is: ‘Out of the individuals whom the model predicted to be earning >50K (TP+FP), how many were classified correctly (TP)?’ This question can be answered by looking at the **Precision** of the model, which is the proportion of positives that are classified correctly: TP/(TP+FP). Another common question is “Out of all the high earning employees with income >50k (TP+FN), how many did the classifier classify correctly (TP)”. This is actually the **Recall**, or the true positive rate: TP/(TP+FN) of the classifier. You might notice that there is an obvious trade-off between precision and recall. For example, given a relatively balanced dataset, a classifier that predicts mostly positive instances, would have a high recall, but a rather low precision as many of the negative instances would be misclassified resulting in a large number of false positives. To see a plot of how these two metrics vary, you can click on the ‘PRECISION/RECALL’ curve in the evaluation result output page (top left part of Figure 7).
 
 ![Binary Classification Evaluation Results](media/machine-learning-evaluate-model-performance/7.png)
+
 Figure 7. Binary Classification Evaluation Results.
 
 Another related metric that is often used is the **F1 Score**, which takes both precision and recall into consideration. It is the harmonic mean of these 2 metrics and is computed as such: F1 = 2 (precision x recall) / (precision + recall). The F1 score is a good way to summarize the evaluation in a single number, but it’s always a good practice to look at both precision and recall together to better understand how a classifier behaves.

articles/machine-learning/machine-learning-troubleshooting-creating-ml-workspace.md

@@ -43,15 +43,15 @@ After the new Machine Learning workspace is created, you can sign in to Machine
 
 **To delete browser cookies**
 
-If you use Internet Explorer, click the **Tools** button in the upper-right corner and select **Internet options**.  
+1. If you use Internet Explorer, click the **Tools** button in the upper-right corner and select **Internet options**.  
 
 ![Internet options][screen4]
 
-Under the **General** tab, click **Delete…**
+2. Under the **General** tab, click **Delete…**
 
 ![General tab][screen5]
 
-In the **Delete Browsing History** dialog box, make sure **Cookies and website data** is selected, and click **Delete**.
+3. In the **Delete Browsing History** dialog box, make sure **Cookies and website data** is selected, and click **Delete**.
 
 ![Delete cookies][screen6]