Hi everyone! Welcome to my blog - there's some other stuff here, like some DevOps IT tutorials and some more personal posts in addition to my main thing - battery posts.
Here I want to begin something I've wanted to give a shot to for a long time - that is, building machine learning for predicting NMC cathode performance. Here's the first chapter and here's the full directory.
I'm not a data scientist by any means; my background/career is in batteries with a lot of experience in web development as well. This project seemed like a good way to get my feet wet with building some machine learning and I figured I'd document my process as well.
For people from the computer science world that know very little about batteries, I've written up two posts, one non-technical and one technical on batteries and cathodes. I think these do a good job of explaining (even for battery people) why a neural net would be very useful for cathode research.
For people from my battery world that know very little about computer science, I highly recommend this series of videos on neural nets from 3Blue1Brown, at least to get started.
For gradient boosting method, I found this video pretty good.
For those from neither of those worlds that are still interested, welcome, and hope the links above can give you enough as well.
Disclaimer
I'm not promising that any of this will work. Rather, I'd like to just document my whole process as I work through the ideas. I will start from the very basics, starting with the simplest possible proof of concept and detail my thinking along each step.