04 - Model Training and Inference¶

Get the Video Course

This section trains a baseline Logistic Regression model for fraud detection and serves predictions via a FastAPI endpoint. The baseline establishes a performance benchmark using sklearn's Pipeline to combine preprocessing (encoding, scaling) with the classifier.

Complete Workflow¶

Inference Workflow¶

Architecture¶

Pre-requisite: Python Environment Setup¶

# Create conda environment
conda create -n mlops-env1 python=3.14 -c conda-forge -y

# Activate environment
conda activate mlops-env1

# Install dependencies (locked versions)
cd ccfd-project
pip install -r requirements.txt

Note: All sections in this course use the same mlops-env1 environment. You only need to create it once. After that, just activate it with conda activate mlops-env1 before running any scripts.

What You Will Learn¶

Demo	Topic	Key Concepts
04_01_Train_Baseline_Model	Train the model	sklearn Pipeline, class_weight='balanced', business metrics
04_02_Inference_API	Serve and test predictions	FastAPI, REST API, browser UI, automated testing

Scripts¶

Script	Purpose	Location
`p3_01_train_model_baseline.py`	Train baseline Logistic Regression	`04_01_Train_Baseline_Model/ccfd-project/`
`p3_02_inference_api_baseline.py`	Serve predictions via FastAPI (port 8000)	`04_02_Inference_API/ccfd-project/`
`p3_03_inference_test_baseline.py`	Test the API with sample transactions	`04_02_Inference_API/ccfd-project/`

Baseline Results Summary¶

Metric	Value	Interpretation
F1 Score	0.2329	Balance of precision/recall
Precision	0.1373	13.7% of fraud predictions are correct
Recall	0.7667	Catches 77% of actual fraud (46/60)
ROC AUC	0.8837	Good discrimination ability
Net Benefit	$39,625	Business value (fraud saved - false alarm costs)

Next Steps¶

Next	Topic	What You'll Do
05_EDA_and_Preprocessing	Exploratory Data Analysis	Analyze fraud patterns, correlations, generate visualizations

Prefer to learn by watching?

The video course builds this whole project with you on screen, step by step.

Get the Video Course

03 - Generate Synthetic Dataset Next: 04 A - Train Baseline Model