Math.NET Numerics
Welcome to our official feedback forum. Do you have an idea? Do you recognize a good idea when you see one? We want to hear from you!
-
89 votes
-
87 votes
-
Provide linux packages (for use with mono)
Extend the package generation scripts and provide the missing documents (and maybe consider to switch back to a classic version system) to build linux packages (e.g. deb for debian/ubuntu) that will run under mono.
46 votes -
Add methods for linear and quadratic programming
Linear and quadratic programming methods are great techniques for minimization problems. Other, more complicated methodologies - such as MPC - can built on top of them fairly easily.
33 votes -
Provide principle component analysis
Extend the linear algebra part with PCA using either Singular Value Decomposition or Eigenvalue Decomposition. Easy implementation, great value!
31 votes -
Implement MATLAB's pchip interpolation algorithm
Piecewise Cubic Hermite Interpolating Polynomial
23 votes -
23 votes
-
Add support for surface interpolation (2D, mesh)
Extend the interpolation methods to support 2D surface interpolation, i.e. on a mesh. See http://community.opensourcedotnet.info/forums/t/565.aspx
23 votes -
Investigate Mono.Simd
The upcoming Mono.Simd assembly is nearing release. Using these functions in certain areas could give massive performance gains.
20 votes -
19 votes
-
VSLab
I think this is a great project. I am suggesting your library for VSLab (http://www.codeplex.com/vslab). I think that an F# wrapper would be great. Are you interested in contribute with your math library in creating a Matlab like environment? Have a look to the project!
19 votes -
implement a real number type that supports exact arithmetic
a real number type that lazily increases its precision when needed up to a maximum user-defined precision.
17 votes -
Add LU and QR decompositions for complex matrices
Currently the LU and QR linear algebra matrix decompositions are only supported on real matrices.
16 votes -
Add row echelon and reduced row echelon form computation
Computing reduced row echelon and row echelon form of a Matrix could be used with Gaussian and Gauss-Jordan elimination techniques.
16 votes -
Release a Silverlight version of Iridium
I just did this. Haven't run the tests yet. Here are my notes:
- Commented all [Serializable] attributes and changed [NonSerialized()] to [IgnoreDataMember]. The default DataContractSerializer will now work correctly.
-Changed Histogram ArrayList buckets to type List<IComparable> and related changes to fields/methods. This maybe should be List<Bucket> as suggested, but the current BinarySearch then needs a IComparer<Bucket>, which doesn't work with the current approach of comparing a Bucket with a double.
- Commented all other classes which utilized ArrayList (they were marked "Obsolete" anyway)
-Commented attributes in AssemblyInfo.cs unsupported in Silverlight
-Changed all calls to System.Array.LongLenth accessors to System.Array.Length
-Removed… more
15 votes -
Solve a system of nonlinear equations
aka fsolve of matlab
12 votes -
Least Square Fit / TDistribution.. What happened to it? Why can't I find them in Numerics lib?
I see these functions existed in old Iridium Release. Are they there in the Numerics Release?
11 votes -
Add stochastic processes
Almost all modern financial analysis are built around stochastic processes. There are many processes, some general like time series and some more specific. It would be very useful for people working in the finance field.
11 votes -
Add single-precision matrices
Memory-bound algorithms could handle larger problems if matrices were stored in single precision (both real and complex). This probably means that single-precision variants of everything else must be supplied too. Lapack, for instance, does this.
10 votes -
Support Array Slicing
Array slicing is very handy. Ideally it would be nice if slicing could be implemented as straight forward as is done python with numpy arrays. However, packages like ublas for C++ also support slicing.
7 votes
