What's new in Al-Kashi 5.0 (release date: Feb 2, 2014)
-------------------------------------------------------
* Support multiple linear regression for 2 independent variables, known issues:
- calculate standard error of regression model parameters: a, b1, and b2
- calculate regression through the origin
* Add basic matrix functions which includes: mAddition, mSubtraction,
mMultiplication, mTranspose, mDeterminants, mCofactor, mAdjoint and mInverse.
* Add "path" function to perform Path analysis to describe the directed dependencies
among a set of variables.
* Add "solve" function to solve a system of equations.
* Kashi will throw now KashiException if there is any error, this enable you to use
try/catch block for more clean object oriented code.
* Fix r-square calculation bug in the regression method.
What's new in Al-Kashi 4.0 (release date: Aug 12, 2013)
-------------------------------------------------------
* Back end function calculates all required information to render Boxplot,
Histogram, Normal Q-Q Plot, and Ternary plot.
* A new "movingAvg" method for moving average which is commonly used with time
series data to smooth out short-term fluctuations and highlight longer-term
trends or cycles.
* Calculate the standard error of regression model parameters, adjusted R square,
significance of regression (i.e. F-statistic and p-value), as well as confidence
intervals at level 95%.
* Add two extra types of mean "Geometric" and "Harmonic" into the mean method.
* Add three extra methods [Jaccard|Dice|Simple Matching] to the "distance" methods in
the "KashiCluster" class to compute the distance in case of binary/quality characters.
* Add two new methods "isSkew" and "isKurt" to find out if skewness or kurtosis is
significant (i.e. test it against 0).
* Add new method "rank" to implement standard competition ranking strategy ("1224" ranking).
* Improve simple linear regression method by add extra origin boolean parameter, if true
it will calculate regression through the origin (i.e. intercept = 0), default is false.
* Add an extra parameter to enable "standardize" method to centers the values around
zero without affect the variance at all (default functionality will make variance=1).
* Add new method "dataLoad" to the "KashiCluster" sub class to simplified using
this functionality.
What's new in Al-Kashi 3.0 (release date: Dec 29, 2012)
-------------------------------------------------------
* Add new method "standardize" to the main Kashi class to perform standardize
transformation (i.e. mean=0 and variance=1).
* Add new Cluster sub class "KashiCluster" where we implement k-means clustering
method which aims to partition n observations into k clusters.
* Implement Hierarchical clustering technique also in the new "KashiCluster" sub
class, currently only Euclidian and Manhattan similarity measures are available
and Average Link as clustering method.
* ANOVA sub class is fully documented now.
What's new in Al-Kashi 2.1 (release date: Sep 15, 2012)
-------------------------------------------------------
* Add new method "inverseNormCDF" for inverse of the standard normal
cumulative distribution, with a given probability.
* Add new method "inverseTCDF" to calculate the t-value of the Student's
t-distribution for a given probability and degrees of freedom.
* Improve the LSD estimation in the ANOVA procedure by using
qt(0.975, error df) x SED instead of 2 x SED
What's new in Al-Kashi 2.0 (release date: Sep 10, 2012)
-------------------------------------------------------
* Add ANOVA sub class "KashiANOVA" which calculate ANOVA table, means,
SE, SED, LSD, and CV values; It is built to be scalable data-analysis
solution and this is why we used MySQL including two loading methods
from array and string (i.e. file content).
Note: Column names ending in an exclamation mark (!) will have this
character removed and the column will be loaded as a factor.
What's new in Al-Kashi 1.2 (release date: Jun 22, 2012)
-------------------------------------------------------
* Add Simpson index to the diversity method.
What's new in Al-Kashi 1.1 (release date: Mar 26, 2012)
-------------------------------------------------------
* This version has no need to connect to MySQL database anymore, thanks to
Pavlos Stamboulides <pavlos@psychology.gr> for the idea.
* The chiDist method returns now exact probability value as float number.
* Add normCDF method to calculate the CDF value for normal distribution.
What's new in Al-Kashi 1.0 (release date: Mar 3, 2012)
-------------------------------------------------------
* First release :o) |