Extensions - Statistical


The com.singularsys.extensions.statistical Extension adds a range of statistical functions.

Aggregate functions

These functions take a set of arguments or an array of data and calculate different statistics.

Name Description Examples
Count Counts its arguments count(1,2,3,4) == 4 count([1,2,3,4])
ElementSumSummation of the arguments sum(1,2,3,4) == 10 sum([1,2,3,4])
MeanMean of the arguments mean(1,2,3,4) == 2.5 mean([1,2,3,4])
MedianMedian of the arguments median(1,2,3,4) == 2.5 median(1,2,3,4,5) == 3
ModeMode of the arguments mode(1,2,3,3,4) == 3
Variance Population/Sample variance and standard deviations of the arguments var_p(1,2,3,4) == 1.25
var_s(1,2,3,4) == 1.67
sd_p(1,2,3,4) == 1.12
sd_s(1,2,3,4) == 1.29
Rank Rank of a single element in a set, default is descending competition rank rank(6,[5,6,7,8]) == 3
Ranks Ranks of all elements in a set, default is descending competition rank ranks([5,6,7,8]) == [4,3,2,1]
ElementMinMax Minimum/maximum of the arguments min(1,2,3,4) == 1 max(1,2,3,4) == 4
ElementSum Sum of the arguments sum(1,2,3,4) == 10
ElementProduct Product of the arguments product(1,2,3,4) == 24

These functions all require a field to be specified and can be added to Jep like other functions. 

Jep jep = new Jep();
FieldI f = new DoubleField();
jep.addFunction("mean", new Mean(f));
jep.reinitializeComponents();
   
Node node = jep.parse("mean(1,2,3,4)");
Object res = jep.evaluate(node);

Some classes like Variance require a flag in the constructor to specify the particular type 

// Standard deviation based on entire population
jep.addFunction("sd", new Variance(Variance.Type.POPSD, f));
Node node2 = jep.parse("sd(1,2,3,4)");
Object res2 = jep.evaluate(node2);

The full setup for the above functions is 

Jep jep = new Jep();
FieldI f = new DoubleField();
jep.addFunction("count", new Count());
jep.addFunction("mean", new Mean(f));
jep.addFunction("median", new Median(f));
jep.addFunction("mode", new Mode(f));
jep.addFunction("mean", new Mean(f));
jep.addFunction("var_p", new Variance(Variance.Type.POPVAR, f));
jep.addFunction("var_s", new Variance(Variance.Type.SAMPLEVAR, f));
jep.addFunction("sd_p", new Variance(Variance.Type.POPSD, f));
jep.addFunction("sd_s", new Variance(Variance.Type.SAMPLESD, f));
jep.addFunction("rank", new Rank(Rank.Type.COMPETITION,true,f));
jep.addFunction("ranks",new Ranks(Ranks.Type.COMPETITION,true,f));
jep.addFunction("min",new ElementMinMax(ElementMinMax.Type.MIN, f));
jep.addFunction("max",new ElementMinMax(ElementMinMax.Type.MAX,f));
jep.addFunction("sum",new ElementSum(f));
jep.addFunction("prod",new ElementProduct(f));
jep.reinitializeComponents();
top

Distributions

A couple of statistical distributions are provided in the com.singularsys.extensions.statistical.distributions package.

Name Description Examples
BinomialDist Binomial distribution. A four argument version is available for Excel compatibility. binomPdf(nSuccess,nTrial,p_success) binomPdf(10,2,0.25)
BinomialCdf Binomial cumulative distribution function binomCdf(nSuccess,nTrial,p_success) binomCdf(10,2,0.25)
NormalDist Normal cumulative distribution function normalCdf(x, mean, sd) normalCdf(0.3, 0, 1)
NormalDist Inverse normal function normalInv(probability,mean,sd) normalInv(0.3 ,0, 1)

These functions would be set up using 

jep.addFunction("binomPdf",new BinomialDist());
jep.addFunction("binomCdf",new BinomialCdf());
jep.addFunction("normalCdf",new NormalDist(NormalDist.Type.CDF));
jep.addFunction("normalInv",new NormalInverse());
top