Streamline complex machine learning projects by designing pipelines that combine preprocessing, modeling, and evaluation. Learn to manage numeric, image, and text data efficiently while preventing data leakage and boosting reproducibility.
Overview
Syllabus
- Scikit-Learn Pipelines
- Master scikit-learn pipelines for efficient data workflows, preprocessing, feature engineering, and model optimization, enhancing machine learning projects with automation and clarity.
- Computer Vision Pipelines
- Explore computer vision by learning image preprocessing, feature extraction, and building classification pipelines with SVMs and CNNs using tools such as OpenCV and PyTorch.
- NLP Pipelines
- Build machine learning pipelines with text data features, including tokenization, vectorization, and part-of-speech tagging with spaCy.
- Project: Data Science Pipeline
- Create a machine learning model pipeline with scikit-learn, using numeric and text data to predict whether or not a customer would recommend a product.
Taught by
Juno Lee (color), Andrew Paster and Arpan Chakraborty