joaovissoci

#Power analysis for CSM

alpha <- 0.05 #alpha level
d <- 29 #degrees of freedom
n <- 185 #sample size
rmsea0 <- 0.08 #null hypothesized RMSEA
rmseaa <- 0.05 #alternative hypothesized RMSEA

#Code below this point need not be changed by user
ncp0 <- (n-1)*d*rmsea0^2

joaovissoci

library(pROC)
data(aSAH)

#Initial ROC analysis
roc(aSAH$outcome, aSAH$s100b) #first argument = outcome; second = predictor
roc(outcome ~ s100b, aSAH) #formula mode with dataset in the end
roc(outcome ~ s100b, aSAH, smooth=TRUE) #applying smoothing

#CI and Plotting
roc1 <- roc(aSAH$outcome,

joaovissoci

#Prediction models in R

#Packages needed:
#"Deducer"
#"tree"
#"randomForest"
#"reshape"

lapply(c("Deduces","randomForest","tree","reshape"), 
library, character.only=T)

joaovissoci

###########################################################################################################
#
# MAXEIG -- performs MAXEIG, but can also be used to perform MAXCOV.
#
# Written by John Ruscio
# Last modified January 9, 2012
#
###########################################################################################################
#

joaovissoci

### ANALISE TAXOMETRICA COM O PACOTE R DE JOHN RUSCIO

# BANCO DE DADOS PARA AS ANALISES
#####
#Preparar um banco com dois ou mais indicadores (algumas an·lises requerem trÍs, como MAXCOV e MAXHEIG), sendo que os "indicadores" 
#s„o escores brutos em escalas ou escores compostos (ponderados ou n„o). Uma alternativa, que pode ajudar na interpretaÁ„o das curvas, È utilizar 
#escores padronizados dos indicadores, embora seja ˙til relatar tambÈm as informaÁıes descritivas da escala original dos indicadores. Se os
#indicadores forem padronizados, lembrar de n„o deixar valores decimais com vÌrgula, que o R n„o reconhece. … necess·rio usar pontos (se n„o
#der certo a substituiÁ„o das vÌrgulas por pontos no Excel, reduzir para duas casas decimais, copiar, colar no TexPad, substituir, colar de volta
#no Excel e salvar). … recomend·vel tambÈm conferir as opÁıes de an·lise oferecidas no manual, uma vez que alguns dos programas do pacote oferecem a opÁ„o de padronizar

joaovissoci

library(‘‘parcor’’)
library(‘‘Matrix’’)
set.seed(100)
adls <- adalasso.net(Data)
network <-
as.matrix(forceSymmetric(adls$pcor.adalasso))
qgraph(network, layout = ‘‘spring’’, labels =
colnames(Data), groups = groups)

joaovissoci

library("IsingFit")
library("qgraph")
library("Matrix")
library("glmnet")
library("IsingSampler")

lapply(c("IsingFit","qgraph","Matrix","glmnet","IsingSampler","psych"), 
library, character.only=T)

joaovissoci

itemfit <- function(x, Zh = TRUE, X2 = FALSE, S_X2 = TRUE, group.size = 150, mincell = 1, S_X2.tables = FALSE,
                    empirical.plot = NULL, empirical.CI = 0, method = 'EAP', Theta = NULL, 
                    impute = 0, ...){
    
    fn <- function(collect, obj, Theta, ...){
        tmpdat <- imputeMissing(obj, Theta)
        tmpmod <- mirt(tmpdat, obj@nfact, pars = vals)
        tmpmod@pars <- obj@pars
        return(itemfit(tmpmod, Theta=Theta, ...))
    } 

joaovissoci

#### USING eRM Package
IRTRolandMorris <- RM(RolandMorrisOriginal)
summary()
coef()
confint(,"beta")
plotI(,i=3)
plotICC(,item.subset=1:4,ask=F,empICC=list("raw"),empCI=list(lty="solid"))
plotPImap()
pp<-person.parameter()
lrt<-LRtest(,se=TRUE)

joaovissoci

#####################################################################################
#TEMPLATE DE ANALISE - contexto_study_3.R
#####################################################################################
#
# Glossário
# $ = Pertence a = data$sexo, que dizer variável sexo pertencente ao banco data
#
#
#####################################################################################
#ORGANIZANDO O AMBIENTE DO R