Richard Ren, NY’s high school student, has designed an air pollution prediction model with 92% accuracy, without any professional training. Ren learned from the free courses available online; used YouTube to learn neural networks. The training data was also found online. He used 3 models – Neural Networks, Random Forest and Logistic Regression – for this project. Interestingly, weather researchers mostly use a single model, limiting their effectiveness.

Click here to read the full story