Sign in
Microsoft.com
 
Microsoft
 
 
 

locked
Multivariate Linear-Gaussian Implementation Issue RRS feed

 > 

  • Question

  • Question
    Sign in to vote
    0
    Sign in to vote


    In infer.NET, is it possible to define a multivariate linear gaussian variable Y so that

    Y = A*X + E

    where X is a multivariate gaussian random variable,

    A is a constant matrix 

    and E is a zero mean multivariate gaussian 

    The following code results in the error below.

    double[,] A_a = { { 4.0, 0.5 }, { 0.5, 7.2 } };
    Matrix A = new Matrix(A_a);

    Variable<Vector> E = Variable.VectorGaussianFromMeanAndVariance(ue, Se).Named("E");

    Variable<Vector> X = Variable.VectorGaussianFromMeanAndPrecision(ux, Kx).Named("X");
    Variable<Vector> Y = A*X + E;

    error:

    operator '*' cannot be applied to operands of type 'Matrix' and 'Variable<vector>'

    Is there perhaps another type of matrix variable that should be use in such a case?

    Monday, July 17, 2017 7:52 AM

All replies

  • Question
    Sign in to vote
    0
    Sign in to vote
    You need to use Variable.MatrixTimesVector.  
    • Proposed as answer by usptact Friday, July 21, 2017 10:24 PM
    Monday, July 17, 2017 6:09 PM
    Owner
  • Question
    Sign in to vote
    0
    Sign in to vote
    Does the matrix A must be observed or declared Constant somehow?
    Tuesday, July 18, 2017 4:56 AM
  • Question
    Sign in to vote
    1
    Sign in to vote
    You can simply pass the Matrix A in your code into Variable.MatrixTimesVector.
    Friday, July 21, 2017 5:46 PM
    Owner