This is a summary of links featured on Quantocracy on Friday, 04/14/2017. To see our most recent links, visit the Quant Mashup. Read on readers!
-
Convolutional Neural Network for Time Series [Quintuitive]Neural networks have been around for a while, but its fair to say that many successful practical applications use at least one convolutional layer. Naturally, convolutions make sense for time series, so I went and added a few to the Walk-Forward Analysis. To make the code easier to use, I ended up creating a self-contained GitHub repository. CNTKs code to create the network layers is
-
Upcoming Webinar: How to use Mixture Models to Predict Market Bottoms w/ @BlackArbsCEO [Quant Insti]The webinar will explain Mixture Models and explore its application to predict an assets return distribution and identify outlier returns that are likely to mean revert. The webinar will cover Why bother? Motivating experimentation with Mixture Models How do Mixture Models work? (An intuitive explanation) Designing the Research Experiment (How do we answer the original question?) Define the