Software Automation
Training Models - Supervised Learning

Supervised Learning

Supervised learning is the most common type of machine learning, where we teach machines by showing them examples with correct answers. Think of it like teaching a child to identify animals by showing them pictures and telling them "this is a cat" or "this is a dog." After seeing enough examples, they can identify new animals they've never seen before.

In supervised learning:

• We provide the algorithm with labelled data (input-output pairs)

• The algorithm learns to map inputs to correct outputs

• Once trained, it can predict outputs for new, unseen inputs

This approach is called "supervised" because we supervise the learning process by providing the correct answers during training. It's like having a teacher who marks your work and tells you exactly what's right or wrong.

Types of Supervised Learning Problems

Supervised learning problems typically fall into two categories:

How Supervised Learning Works

The process of supervised learning involves these key steps:

1. Data Collection: Gather relevant data with both features and corresponding labels

2. Data Preparation: Clean the data, handle missing values, and split into training and testing sets

3. Model Selection: Choose an appropriate algorithm based on the problem type

4. Training: Feed the training data to the algorithm so it can learn patterns

5. Evaluation: Test the model on unseen data to assess its performance

6. Deployment: Use the trained model to make predictions on new data

Interactive Activity: Teachable Machine

Let's get hands-on with supervised learning using Google's Teachable Machine, a tool that allows you to create your own image classification models without writing any code.

Activity Instructions:

1. Visit Teachable Machine

2. Create a new image project

3. Set up at least 3 classes (e.g., "Thumbs Up", "Thumbs Down", "Peace Sign")

4. For each class, collect at least 20 samples by using your webcam

5. Train your model (click the "Train Model" button)

6. Test your model by performing gestures and seeing if it correctly classifies them

7. Experiment with:

   • Adding more samples to improve accuracy

   • Testing in different lighting conditions

   • Adding a new class and re-training

Advantages and Limitations

Common Supervised Learning Algorithms

• Linear Regression: Predicts continuous values using a linear approach

• Logistic Regression: Used for binary classification problems

• Decision Trees: Uses a tree-like model of decisions

• Random Forest: Combines multiple decision trees for better accuracy

• Support Vector Machines (SVM): Finds the optimal boundary between classes

• K-Nearest Neighbors (KNN): Classifies based on similarity to known examples

• Neural Networks: Mimics the human brain to recognize patterns

Activity: Identify the Problem Type

For each scenario, identify whether it's a classification or regression problem, and explain what features and labels would be used:

1. Predicting whether a student will pass or fail a course based on attendance, homework completion, and previous grades

2. Estimating how many customers will visit a store on a given day

3. Determining if a credit card transaction is fraudulent

4. Predicting the growth height of a plant based on sunlight, water, and fertilizer

5. Categorizing news articles as sports, politics, entertainment, or technology