Space Science with Python - AI 1-5: Support Vector Machines - Intro
GitHub Link of today's session: https://github.com/ThomasAlbin/Astron...
-- Space Science with Python - AI 1-5: Support Vector Machines - Intro --
Machine Learning and advanced analytics do not always require state-of-the-art neural networks and deep learning architectures. Sometimes, "non-neural-network" solutions may perform well, too.
In this first ML approach, we are going to distinguish between X-class asteroid spectra and Non-X-classes (so we reduce a multi-class problem to some introductory binary problem).
This tutorial session uses scikit-learn and the Support Vector Machine implementation to achieve remarkable results by applying this method on the naive data.
Next session: Optimizing / Fine Tuning the best parameters for Support Vector Machines.
---
How to setup a local dev environment: • Space Science with Python - Part 2: S...
Machine Learning Mastery on Metrics: https://machinelearningmastery.com/to...
https://machinelearningmastery.com/wp...
---
Content
0:00 Introduction + GitHub
2:45 Theory of Support Vector Machines
8:44 Python - Data Preparation
13:18 Python - Training / Validation / Test Data
20:52 Python - Scaling
23:30 Python - Support Vector Machine Training
25:42 Python - Check Test Data / Metrics
37:00 Python - Naive Baseline
39:37 Summary
---
There is a lot to do and to learn and I hope you will join the journey. Meanwhile, if you have questions or ideas, reach out to me via:
Twitter: / mrastrothomas
Reddit: / mrastrothomas
GitHub: https://github.com/ThomasAlbin
Or drop a comment!
Talk to you later,
Thomas