Machine Learning Directed Study: Report 1
This document is a report on a machine learning directed study project focused on analyzing satellite debris. The author gathered data by downloading 3D models of a CubeSat from GrabCad. They used CAD software to derive fundamental properties from each part of the satellite and automated the process using an AutoHotkey script. The data was then collected and converted into a .csv file for easy import into Matlab. The properties considered for analysis were mass, volume, density, area, bounding box volume, and material. The author used a Splom plot to visualize the data and found that most properties correlated with each other. They attempted to perform PCA but found that the dataset was too small and had high variation. Next, they used k-means clustering on the data, focusing on volume and mass as critical columns. The clusters were analyzed and plotted, revealing the need for more data in Cluster 3. The author concludes by stating that the dataset needs to be expanded in terms of both quantity and variety of data, and that more advanced analysis, such as PCA, will be attempted in the future.
This content was originally posted on my projects website here. The above summary was made by the Kagi Summarizer