This is a method added to the vows API, but hasn't made it to the documentation yet.
Add function to calculate the percentile rank of a value in an array. Signed-off-by: Trevor Norris <firstname.lastname@example.org>
There is a lot more to do, but made mention of the upcoming changes in the README, and finished core.js.
The files are small enough, and is more convenient for users that don't want to have to build from scratch.
Quantile calculation now more closely follows the scipy implementation. Added further test and documentation. Refactored test helper functions. Signed-off-by: Trevor Norris <email@example.com>
Signed-off-by: Trevor Norris <firstname.lastname@example.org>
Use the log form of the Beta function for generating the PDF for the Beta distribution to avoid NaNs for large parameter values. Signed-off-by: Trevor Norris <email@example.com>
Fix issue #73, 'What types of correlation does jstat do?' Clarify Documentation: - Which standard deviation estimator is calculated by stdev() - Which variance estimator is calculated by variance() - Which correlation coefficient is calculated by corrcoeff() (Pearson's Rho) Signed-off-by: Trevor Norris <firstname.lastname@example.org>
On systems with node installed as both 'node' and 'nodejs', 'make' will evaluate 'which node nodejs' to include both the path to 'node' and the path to 'nodejs', ultimatley psasing 'nodejs' as the first argument to 'node' and throwing an error. In order to fix this, and make 'make doc' run on systems with node aliased to both 'node' and 'nodejs', I just added a 'grep' to select the first of 'node' and 'nodejs' if both are defined. Tested on Ubuntu 13.10 with GNU Make 3.81, GNU bash 4.2.45(1)-release and Node.js v0.10.26. Signed-off-by: Trevor Norris <email@example.com>
What was I thinking using so many spaces with those parens?
Didn't belong at the top of core, and license file was out of date.
Updates for the Regression Module
Fixed some issues with missing semicolons Updated regresst function (anova results, Standard Error, T score & p value for regressors) Removed the usage of function wide variables
Error in that cumsum forgot to include the default return statement for all other possible cases.
cumsum() fixed for matrices
Adding information on how the community can communicate with us.