AutoML Model Training with MLJAR 🚀🤖

In this notebook, you’ll navigate four key steps to take your data from raw file to trained Meachine Learning models. MLJAR's AutoML engine will automatically test multiple algorithms, optimize hyperparameters, and generate interactive report for you.

Steps in this notebook:

1. Load your dataset
2. Define features and target
3. Train with MLJAR AutoML
4. Review the comprehensive report

All required packages will be installed when running recipes. Let’s get started!

Step 1 – Load Your Dataset 📥

Bring your data into the notebook easily:

• 🍰 Recipe: Read CSV — Read a CSV file with UI Code Recipe.
• 🤖 AI Read File — AI auto-detects file type and loads your file into a DataFrame named `df`.

After loading, you’ll see the first few rows to verify everything loaded correctly.

print('🍰 Recipe: Read CSV   or   🤖 AI Read File')

Step 2 – Define Features (X) & Target (y) 🎯

Prepare your modeling data:

• 🔧 Select X, y — Use Code Recipe UI to select X and y.

Once done, you’ll have `X` (predictors) and `y` (outcome) ready for training.

print('🍰 Recipe: Select X, y')

Step 3 – Train with MLJAR AutoML 🤖⚙️

Launch automated model training:

• Select Features (X) and Target (y) in the AutoML recipe UI.
• Execute code to start MLJAR AutoML training.
• Wait a while ☕ and watch logs as multiple models are trained and validated under the hood.

print('🍰 Recipe: Train AutoML')

Step 4 – Review AutoML Report 📑✨

Explore an interactive report. You can click on models in the leadorboard table to get details for each of them.

print('🍰 Recipe: View AutoML Report')

Next steps

Your AutoML training is done. You have the best model ready. You can use it for computing predictions on new data. It can be done in many ways.

• You can have batch script that is computing predicions for new data and insert into database.
• You can deploy your model as REST API, and send new data points in requests.
• You can create a web application and build UI for your Machine Learning model.

If you have questions or need help, please just let me know at piotr@mljar.com - I will help you :)