gamma_index - calculation of gamma index on multi-dimensional distributions
Project description
Small Python library for calculation of gamma index.
The module offers two functions:
- gamma_matrix
This function calculates gamma index in each point for two n-dimensional distributions (with a same shape). It is not very optimized, it’s O(n x n), where n is the total number of elements of the matrix. It takes minutes to calculate the 3-D matrix for 50x50x50 distribution. Time scales roughly as n^6 (n = size of square matrix in one dimension)
- gamma_matrix_c
This functiion calculates gamma index using a slightly optimized C algorithm. It becomes more efficient than gamma_matrix with a matrix size of roughly 1000 elements for default parameters. However, the relative efficiency of this function highly depends on the parameters and values (the better agreement, the faster the algorithm is). Time scales roughly as n^3.2 (n = size of square matrix in one dimension)
- gamma_matrix_pass
This function returns a boolean matrix (passed / not passed) with optimized search only on a narrow neighbourhood of each point.
TODO
Prepare an optimized version of gamma_matrix based on layers
Add tests
