module 5

_quarto.yml

@@ -127,6 +127,25 @@ website:
           - text: '   7.1. Exercises'
             href: modules/module4/module4-25-testing_your_svm_rbf_knowledge.qmd
           - href: modules/module4/module4-28-what_did_we_just_learn.qmd    
+      - section: "**M5. Preprocessing Numerical Features, Pipelines and Hyperparameter Optimization**"
+        contents:
+          - href: modules/module5/module5-00-module_learning_outcomes.qmd
+          - href: modules/module5/module5-01-the_importance_of_preprocessing.qmd
+          - text: '   1.1. Exercises'
+            href: modules/module5/module5-02-questions_on_why.qmd
+          - href: modules/module5/module5-05-case_study:_preprocessing_with_imputation.qmd
+          - text: '   2.1. Exercises'
+            href: modules/module5/module5-06-imputation.qmd
+          - href: modules/module5/module5-09-case_study:_preprocessing_with_scaling.qmd
+          - text: '   3.1. Exercises'
+            href: modules/module5/module5-10-name_that_scaling_method.qmd
+          - href: modules/module5/module5-13-case_study:_pipelines.qmd
+          - text: '   4.1. Exercises'
+            href: modules/module5/module5-14-pipeline_questions.qmd
+          - href: modules/module5/module5-17-automated_hyperparameter_optimization.qmd
+          - text: '   5.1. Exercises'
+            href: modules/module5/module5-18-exhaustive_or_randomized_grid_search.qmd
+          - href: modules/module5/module5-21-what_did_we_just_learn.qmd
 
 # Since we are declaring options for two formats here (html and revealjs)
 # each qmd file needs to include a yaml block including which format to use for that file.

modules/module5/module5-00-module_learning_outcomes.qmd

@@ -0,0 +1,29 @@
+---
+format:
+  html:
+    page-layout: full
+---
+
+# 0. Module Learning Outcomes
+
+::: {.panel-tabset .nav-pills}
+
+## Video
+
+<iframe
+    class="video"
+    src="https://www.youtube.com/embed/9zkvbdTvS6w" 
+    title="Module 5 Video - Module Learning Outcomes"
+    frameborder="0"
+    allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture"
+    allowfullscreen
+></iframe>
+
+## Slides
+
+<iframe
+    class="slide-deck"
+    src="slides/module5_00.html"
+></iframe>
+
+:::

modules/module5/module5-01-the_importance_of_preprocessing.qmd

@@ -0,0 +1,29 @@
+---
+format:
+  html:
+    page-layout: full
+---
+
+# 1. The Importance of Preprocessing
+
+::: {.panel-tabset .nav-pills}
+
+## Video
+
+<iframe
+    class="video"
+    src="https://www.youtube.com/embed/aBpgJh0ikoA?start=2473&end=3131&rel=0" 
+    title="Module 5 Video - The Importance of Preprocessing"
+    frameborder="0"
+    allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture"
+    allowfullscreen
+></iframe>
+
+## Slides
+
+<iframe
+    class="slide-deck"
+    src="slides/module5_01.html"
+></iframe>
+
+:::

modules/module5/module5-02-questions_on_why.qmd

@@ -0,0 +1,104 @@
+---
+format: html
+---
+
+<script src='../../src/quiz.js'></script>
+
+# 1.1. Exercises
+
+## Questions on Why
+
+<div id='mcq1'></div>
+<script>
+    generateQuiz(
+        'mcq1',
+        'Question 1',
+        'Which model will still produce meaningful predictions with different scaled column values?',
+        {
+        'Decision Trees': 'You are right! Decision Trees visit a single feature at a time, unlike 𝑘-NN models, which calculate distances using the features all together.',
+        '𝑘-NN': 'This works in a similar way to SVM, where distance is calculated and features are observed together and not independently of each other.',
+        'SVM': 'This works in a similar way to 𝑘-NN, where distance is calculated and features are observed together and not independently of each other.',
+        },
+        'Decision Trees',
+    );
+</script>
+
+<div id='mcq2'></div>
+<script>
+    generateQuiz(
+        'mcq2',
+        'Complete the following statement',
+        'Preprocessing is done ____.',
+        {
+        'To the model before training': 'It is done before training but not to the model.',
+        'To the data before training the model': '',
+        'To the model after training': 'It’s not done to the model or after training.',
+        'To the data after training the model': 'You are half right, but it’s not done after training the model.',
+        },
+        'To the data before training the model',
+    );
+</script>
+
+## Motivation True and False
+
+<div id='mcq3'></div>
+<script>
+    generateQuiz(
+        'mcq3',
+        'Question 1',
+        'Columns with lower magnitudes compared to columns with higher magnitudes are less important when making predictions.',
+        {
+        'True': 'These types of models find examples that are most similar to the text example in the <i>training</i> set.',
+        'False': 'Great! Just because a feature has smaller values does not mean it’s less informative.',
+        },
+        'False',
+    );
+</script>
+
+<div id='mcq4'></div>
+<script>
+    generateQuiz(
+        'mcq4',
+        'Question 2',
+        'A model less sensitive to the scale of the data makes it more robust.',
+        {
+        'True': 'Nailed it!',
+        'False': 'Models that are more sensitive to scale can be problematic.',
+        },
+        'True',
+    );
+</script>
+
+## Preprocessing Questions
+
+<div id='mcq5'></div>
+<script>
+    generateQuiz(
+        'mcq5',
+        'Question 1',
+        '<code>StandardScaler</code> is a type of what?',
+        {
+        'Converter': 'You are close but the terminology is not correct.',
+        'Categorizer': 'Not quite.',
+        'Model': 'We are not modeling with <code>StandardScaler</code> since it is <i>transforming</i> the data and not predicting on it.',
+        'Transformer': '',
+        },
+        'Transformer',
+    );
+</script>
+
+<div id='mcq6'></div>
+<script>
+    generateQuiz(
+        'mcq6',
+        'Question 2',
+        'What data does <code>StandardScaler</code> alter?',
+        {
+        'Training only': 'It definitely alters the training set but that’s not all.',
+        'Testing only': 'It does alter the test data but it doesn’t stop there.',
+        'Both training and testing': '',
+        'Neither training or testing': 'We need to scale our data though!',
+        },
+        'Both training and testing',
+    );
+</script>

modules/module5/module5-05-case_study:_preprocessing_with_imputation.qmd

@@ -0,0 +1,29 @@
+---
+format:
+  html:
+    page-layout: full
+---
+
+# 2. Case Study: Preprocessing with Imputation
+
+::: {.panel-tabset .nav-pills}
+
+## Video
+
+<iframe
+    class="video"
+    src="https://www.youtube.com/embed/aBpgJh0ikoA?start=3137&end=3722&rel=0" 
+    title="Module 5 Video - Case Study: Preprocessing with Imputation"
+    frameborder="0"
+    allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture"
+    allowfullscreen
+></iframe>
+
+## Slides
+
+<iframe
+    class="slide-deck"
+    src="slides/module5_05.html"
+></iframe>
+
+:::