Fast Matrix for Java (fm4j) is a general-purpose matrix utility library for computing with dense matrices.
fm4j encapsulated different underlying implementations and select the optimal one in run-time depending on the size of the input matrix. Moreover, fm4j employs Java (Tm) Concurrency to take advantage of the computation power of multi-cor processors.
- Basic operations: Addition, Subtraction, Multiplication, Scalar Multiplication
- Matrix Properties: Trace, Determinant, Transpose, Inverse, Eigenvalues, Singular values, Mean and Variance, Covariance
- System of Linear Equations: Solve Full-Ranked system, Linear Regression
- Decompositions: Cholesky Decomposition, LU Decomposition, QR Decomposition, Hessenberg Decomposition, Schur Decomposition, SVD
- Signal Analysis: (Inverse) Discrete Fourier Transform, (Inverse) Discrete Fourier Transform in 2-D, (Inverse) Discrete Fourier Transform in Hyper-dimension, (Inverse) Discrete Haar Wavelet Transform
- Clustering: K-Mean, Gaussian Mixture Model, DBSCAN
- Spatial Functions: Implicit kd-tree, k-Nearest-Neighbor search, Query Window search
- Classification: Logistic Regression, Feed-Forward Back-Propagate Neural Network
- Special Matrix Construction: Null matrix, Identity matrix, Vandermonde matrix, Givens rotational matrix, Hilbert matrix, Augmented matrix, Kernel Trick
- Import and Export: Serialization to text format, Serialization to code format, Read a matrix from a stream of number, Read a matrix from code format
Be the first to post a review of Fast Matrix for Java!