💻 Week 09 - Class Roadmap (90 min)

2022/23 Lent Term

📚 Learning Objectives

You will learn how to:

• Create a confusion matrix
• Calculate the accuracy of a model
• Calculate the precision and recall of a model
• Judge the performance of a ML models

Note: we will be talking about classification, a supervised learning approach. This means that we will be using a dataset with a response variable (also known as the target variable) to predict the response variable (a class) for new observations.

🛣️ Roadmap

⚙️ Setup (~ 10 min)

🎯 ACTION POINTS:

1. Go to Moodle and download the files for this week’s lab.
2. Open RStudio and create a new project.
3. Add the files you download to the folder of your project.
4. Open DS101L_2022_23_W09_lab.Rmd in RStudio.
5. Run the code chunks in the ⚙️ Setup section to load the libraries you need for this lab.

Ask your colleagues and tutor for help if you get stuck.

1️⃣ Load Data (~ 15 min)

🧑‍🏫 INSTRUCTOR NOTES:

1. Your instructor will explain what is in the data.
2. Since we will take a supervised learning approach, your instructor will briefly explain what the input and output variables are.

🎯 ACTION POINTS:

Follow the action points in the markdown file.

2️⃣ Training vs Test (~35 min)

👨‍🏫 TEACHING POINT:

1. Your instructor will tell you about a common strategy to validate Machine Learning models: training versus test splits.

🎯 ACTION POINTS:

Follow the action points in the markdown file. You will learn how to create a training and test split and what a confusion matrix looks like.

There are also several points of discussion in the markdown file. You will discuss these points with your colleagues and tutor.

3️⃣ Metrics (~ 30 min)

👨‍🏫 TEACHING POINT:

1. Each cell in the confusion matrix can be identified with a name. Your instructor will tell you about:

   • False Positives
   • False Negatives
   • True Positives
   • True Negatives

2. Take a look at this table of metrics together.

👥 WORKING TOGETHER IN PAIRS:

Try to work out the solution to the questions posed in the markdown file. You can discuss these questions with your colleagues and tutor.

👨‍🏫 TEACHING POINT:

1. Your instructor will explain what is good and bad about the metrics of this classification model!

🏡Take-home exercise

There is a take-home exercise in the markdown file. Can you answer the questions posed in there?

If you can’t work out the answers to the questions, send us a message on Slack!