Elements of effective machine learning datasets in astronomy by Bernie Boscoe et al. on Tuesday 29 NovemberIn this work, we identify elements of effective machine learning datasets inastronomy and present suggestions for their design and creation. Machinelearning has become an increasingly important tool for analyzing andunderstanding the large-scale flood of data in astronomy. To take advantage ofthese tools, datasets are required for training and testing. However, buildingmachine learning datasets for astronomy can be challenging. Astronomical datais collected from instruments built to explore science questions in atraditional fashion rather than to conduct machine learning. Thus, it is oftenthe case that raw data, or even downstream processed data is not in a formamenable to machine learning. We explore the construction of machine learningdatasets and we ask: what elements define effective machine learning datasets?We define effective machine learning datasets in astronomy to be formed withwell-defined data points, structure, and metadata. We discuss why theseelements are important for astronomical applications and ways to put them inpractice. We posit that these qualities not only make the data suitable formachine learning, they also help to foster usable, reusable, and replicablescience practices.arXiv: http://arxiv.org/abs/http://arxiv.org/abs/2211.14401v1