A neural network built from basic linear algebra coded in C# trained on a regression model for stock prices as a colaborative project for B.C. Highschool Capstone requirement 2019/2020. It is specifically done without any pre-existing machine learning api so that the we can have a greater understanding of the structure of the neural network and stocastic gradient descent. Achieved a minimum of 3% loss on one stock with about 200 training data.
Presentation Video: https://www.youtube.com/watch?v=6bzkh_sOCoE&feature=youtu.be