Supervised Learning
Supervised learning is the class/subcategory of machine learning applications that uses labeled datasets to train algorithms to classify data or predict outcomes. In effect, a desired output (a supervisory signal) is provided to train a model on input objects. The overarching forms that supervised learning take are: classification or regression tasks.
The following sections provide summaries and links to relevant pedagogical reference material as well as algorithm use cases focusing primarily on astrobiology and exoplanetary science tasks. Some of these materials will be cross-linked to the Project Repository page.
Reference Materials:
Explanation of supervised learning (video).
Overview of supervised learning and Python example of Logistic Regression (classification).
Discussion of supervised learning and example of Support Vector classification.
Discussion and tutorial using Keras and Tensorflow.
Use Cases:
· Exoplanet transit detection use case: Example includes Neural Network and Support Vector Machine supervised learning models.
Exoplanet hunting with Deep Learning by Gabriel Garza. Description and guide available here. GitHub repository with code and data available here