This is a summary of links featured on Quantocracy on Wednesday, 07/18/2018. To see our most recent links, visit the Quant Mashup. Read on readers!
Stock Prediction with ML: Walk-forward Modeling [Alpha Scientist]Key Takeaways: Traditional methods of validation and cross-validation are problematic for time series prediction problems The solution is to use a "walk-forward" approach which incorporates new information as it becomes available. This approach gives us a more realistic view of how effective our model would truly have been in the past, and helps to avoid the overfitting trap. It's
Our Conversation with Tobias Carlisle (@Greenbackd) [Flirting with Models]This post covers our conversation with Tobias Carlisle, which you can listen to here. 2:09 – Toby starts at the beginning: with school classes that included sheering sheep in Australia. Corey Hoffstein ("CH"): I was so taken aback by this introduction that I was totally caught off-guard. I knew Toby had grown up in a fairly remote town in Australia (he likes to joke he's the only
10 Reasons for loving Nearest Neighbors algorithm [Quant Dare]I fell in love with k-Nearest Neighbors algorithm at first sight, but it isnt blind love. I have plenty of reasons to be mad about it. 1. Its pretty intuitive and simple Given that all you need to do is to compare samples, the Nearest Neighbors (k-NN) algorithm is a perfect first step to introduce Machine Learning. Its very simple to understand, easy to explain and perfect to demonstrate