A repository of machine learning online course delivered by DataTalks.Club. More about the course is available on this page and github link.
Breakdown structure of this course into two parts:
- Machine learning algorithms: Linear Regression, Classification, Trees (Decision Tree and Ensembles such as Random Forest and XGBoost), and Deep Learning. Python, Scikit-learn, Keras and TensorFlow are main libraries in this development stage.
- Bringing machine learning into practical applications to solve problems in real-world scenarios with Flask, FastAPI, BentoML, Docker and Kubernetes.
- Course introduction, history and evolution.
- Basics of programming for ML (Python).
- Understanding the basics of data science & machine learning.
- Exploratory Data Analysis techniques.
- Fundamentals of linear regression.
- Implementation of simple linear regression model using Python libraries like NumPy and Scikit-Learn.
- Basic level of feature engineering.
- Advanced topics in linear regression including regularization and polynomial regression, logistic regression etc.
Week3: Classification
- Overview of logistic regression.
- Implementing a basic logistic regression algorithm in Python.
- Handling categorical variables with Encoding.
- Feature Importance.
- Representative data for better performance: Cross validation.
- Confusion Table for Precision and Recall.
- Measuring performance trade-off with ROC curves.
Week5: Machine Learning Deployment
- Introduction of Web Service with Flask.
- Bringing machine learning from notebooks to services and applications with Flask.
- Dependency management with pipenv.
- Package Flask applications with machine learning models and their dependencies with Docker.
- More interactive model serving with FastAPI.
More to be updated soon.
2023 Update: The section "Production with BentoML" has been removed from the source repository. However, you can explore my work utilizing BentoML on this directory