| Title: | Fit Univariate Mixed and Usual Distributions |
|---|---|
| Description: | Extends the fitdist() (from 'fitdistrplus') adding the Anderson-Darling ad.test() (from 'ADGofTest') and Kolmogorov Smirnov Test ks.test() inside, trying the distributions from 'stats' package by default and offering a second function which uses mixed distributions to fit, this distributions are split with unsupervised learning, with Mclust() function (from 'mclust'). |
| Authors: | José Carlos Del Valle <[email protected]> |
| Maintainer: | José Carlos Del Valle <[email protected]> |
| License: | GPL-3 |
| Version: | 3.1.0 |
| Built: | 2024-11-27 05:55:47 UTC |
| Source: | https://github.com/jcval94/fitultd |
Fit of univariate distributions with censored data ignored by default or can be inputed.
FDist(X, gen = 1, Cont = TRUE, inputNA, plot = FALSE, p.val_min = 0.05, crit = 2, DPQR = TRUE)FDist(X, gen = 1, Cont = TRUE, inputNA, plot = FALSE, p.val_min = 0.05, crit = 2, DPQR = TRUE)
X |
A random sample to be fitted. |
gen |
A positive integer, indicates the sample length to be generated by the fit, 1 by default. |
Cont |
TRUE, by default the distribution is considered as continuos. |
inputNA |
A number to replace censored values, if is missing, only non-censored values will be evaluated. |
plot |
FALSE. If TRUE, a plot showing the data distribution will be given. |
p.val_min |
0.05, minimum p.value for Anderson Darling and KS Test to non-reject the null hypothesis and continue with the process. |
crit |
A positive integer to define which test will use. If 1, show the distributions which were non-rejected by the Anderson Darling or Kolmogorov Smirnov tests, in other cases the criterion is that they mustn't be rejected by both tests. |
DPQR |
TRUE, creates the distribution function, density and quantile function with the names dfit, pfit and qfit. |
Calculate the distribution name with parameters, a function to reproduce random values from that distribution, a numeric vector of random numbers from that function, Anderson Darling and KS p.values, a plot showing the distribution difference between the real sample and the generated values and a list with the random deviates genetator, the distribution function, density and quantile function
set.seed(31109) FIT1<-FDist(rnorm(1000,10),gen=100,p.val_min=.03,crit=1,plot=TRUE) #Random Variable FIT1[[1]] #Random numbers generator FIT1[[2]]() #Random sample FIT1[[3]] #Goodness of fit tests results FIT1[[4]] #Plot FIT1[[5]] #Functions r, p, d, q FIT1[[6]]set.seed(31109) FIT1<-FDist(rnorm(1000,10),gen=100,p.val_min=.03,crit=1,plot=TRUE) #Random Variable FIT1[[1]] #Random numbers generator FIT1[[2]]() #Random sample FIT1[[3]] #Goodness of fit tests results FIT1[[4]] #Plot FIT1[[5]] #Functions r, p, d, q FIT1[[6]]
Fits a set of observations (random variable) to test whether is drawn from a certain distribution
FDistUlt(X, n.obs = length(X), ref = "OP", crt = 1, plot = FALSE, subplot = FALSE, p.val_min = 0.05)FDistUlt(X, n.obs = length(X), ref = "OP", crt = 1, plot = FALSE, subplot = FALSE, p.val_min = 0.05)
X |
A random sample to be fitted. |
n.obs |
A positive integer, is the length of the random sample to be generated |
ref |
Aumber of clusters to use by the kmeans function to split the distribution, if isn't a number, uses mclust classification by default. |
crt |
Criteria to be given to FDist() function |
plot |
FALSE. If TRUE, generates a plot of the density function. |
subplot |
FALSE. If TRUE, generates the plot of the mixed density function's partitions. |
p.val_min |
Minimum p.value to be given to non-reject the null hypothesis. |
A list with the density functions, a random sample, a data frame with the KS and AD p.values results, the corresponding plots, random numbers generator functions and a table with all other possible distributions
set.seed(3110934) X<-c(rnorm(493,189,12),rweibull(182,401,87),rgamma(190,40,19)) A_X<-FDistUlt(X,plot=TRUE,subplot=TRUE) A_X<-FDistUlt(X,plot=TRUE,subplot=TRUE,p.val_min=.005) # Functions generated A_X[[1]][[1]]() # Random sample A_X[[2]] #Distributions A_X[[3]] # Plots par(mfrow=c(1,2)) A_X[[4]][[1]] A_X[[4]][[2]] # More functions A_X[[5]][[1]]()set.seed(3110934) X<-c(rnorm(493,189,12),rweibull(182,401,87),rgamma(190,40,19)) A_X<-FDistUlt(X,plot=TRUE,subplot=TRUE) A_X<-FDistUlt(X,plot=TRUE,subplot=TRUE,p.val_min=.005) # Functions generated A_X[[1]][[1]]() # Random sample A_X[[2]] #Distributions A_X[[3]] # Plots par(mfrow=c(1,2)) A_X[[4]][[1]] A_X[[4]][[2]] # More functions A_X[[5]][[1]]()