brisk 0.1

Fast implementation of numerical functions using Numba

Optimized numerical computation using Continuum’s Numba. Intended as a drop-in replacement for numerical functions in numpy, scipy, or builtins. Provides strong performance boosts.

Numba website

Inputs use numpy arrays, not lists. Rough/early release - Open to suggestions and bug reports.

Included functions

• sum: Similar to builtin sum, or numpy.sum

• mean: Similar to numpy.mean

• var: Variance test, similar to numpy.var

• cov: Covariance estimation, similar to numpy.cov

• std: Standard deviation, similar to numpy.std

• corr: Pearson correlation test, similar to scipy.stats.pearsonr

• bisect: Similar to standard library bisect.bisect

• bisect_left: Similar to standard library builtin.bisect_left

• interp: Linear interpoliation, similar to numpy.interp. x is an array.

• interp_one: Linear interpolation, similar to numpy.interp. x is a single value.

• detrend: Similar to scipy.signal.detrend. Linear or constant trend.

• ols: Simple Ordinary Least Squares regression for two data sets.

• ols_single: Simple Ordinary Least Squares regression for one data set.

• lin_resids: Residuals calculation from a linear regression with two data sets

• lin_resids_single: Residuals calculation from a linear regression with one data set.