New Post: Negative Samples from Geometric Distribution

Christoph,

Thanks for getting to this so quickly!

I really like the design of your library. And it is extremely useful for my purposes.

I am designing custom compression algorithms for a specific class of time series data (Brownian Motion / Random Walk).

I would like to generalize the algorithms (as much as possible) for all .NET primitive numeric types:

• Signed and Unsigned Integral types (perhaps even including BigInteger),
• Single, Double,, and Decimal.
• DateTime and TimeSpan (Ticks).
  

Primarily I am focused on financial market data (I work for a hedge fund). But the algorithms seem to be very beneficial for use in a variety of other domains.

One of the features I would like to see added to your "Generate" static method implementations, is the ability to overload with a "Granularity" argument. I am building this capability into my own wrappers, but many others would probably find it quite useful.

The other thing that I am building as a part of my project is a Data Modeler application (using WPF MVVM with OxyPlot). This will allow me to easily simulate various types of time series data with specific characteristics (think Monte Carlo). If this turns out well, and if it seems like something that contributors on your project might like to extend or improve, I'd be happy to donate the code and help maintain it. I'll probably design it as composite application based on Prism, so that it can benefit from a "plug-in" architecture.

I actually have more "real" sample data than I know what to do with. But when it comes to generalizing compression algorithms, the more the merrier!

Again, thanks for your great work on this project!

BTW: There is an interesting blog post by Bart De Meyer (async-await, etc.) concerning OxyPlot that other MathNet.Numerics users might find helpful:
Creating Graphs in WPF using OxyPlot