In terminological/tidy-info-stats: Functions for manipulating information statistics

knitr::opts_chunk$set(echo = TRUE)

library(infotheo)
library(tidyverse)
library(ggplot2)
library(devtools)
#library(tidyinfostats)
devtools::load_all("~/Git/tidy-info-stats/")
set.seed(101)
theme_set(standardPrintOutput::defaultFigureLayout())

Check entropy calculations working:

Assume a continuous distribution

hb = ConditionalDistribution$new()
hb$withDistribution(0.75, LogNormalDistribution$new(mode=12,var=2), "asymptomatic")
hb$withDistribution(0.25, LogNormalDistribution$new(mode=8,var=3), "tired")
hb$withDistribution(0, LogNormalDistribution$new(mode=4,var=5), "unwell")


k = ConditionalDistribution$new()
k$withDistribution(0.4, NormalDistribution$new(mean=1,sd=0.5), "unwell")
k$withDistribution(0.6, NormalDistribution$new(mean=2,sd=1), "asymptomatic")
k$withDistribution(0, NormalDistribution$new(mean=8,sd=3), "tired")

na = ConditionalDistribution$new()
na$withDistribution(0, NormalDistribution$new(mean=15,sd=5), "unwell")
na$withDistribution(1, NormalDistribution$new(mean=14,sd=5), "asymptomatic")
na$withDistribution(0, NormalDistribution$new(mean=12,sd=5), "tired")
na$plot(100,200)

tibble(measure = c("Mean","Variance","Mutual Information"),
hb = c(hb$theoreticalMean(),hb$theoreticalVariance(), hb$theoreticalMI()),
k = c(k$theoreticalMean(),k$theoreticalVariance(),k$theoreticalMI()),
na = c(na$theoreticalMean(),na$theoreticalVariance(),na$theoreticalMI()))

testData = k$sample(1000) %>% mutate(test="k") %>% bind_rows(
        hb$sample(500) %>% mutate(test="hb")) %>% bind_rows(
        na$sample(100) %>% mutate(test="na")
    ) %>% rename(outcome=y, value=x)
testData = testData[sample(nrow(testData)),]

testData %>% group_by(test) %>% calculateDiscreteContinuousMI(vars(outcome), value, method = "KWindow")
testData %>% group_by(test) %>% summariseObservations(vars(outcome))
missingTest = testData %>% group_by(test) %>% expectedObservations(vars(outcome),600)
devtools::load_all("~/Git/tidy-info-stats/")
testData %>% group_by(test) %>% calculateDiscreteContinuousUnlabelledMI(vars(outcome), value, expectedCount=missingTest)

devtools::load_all("~/Git/tidy-info-stats/")
#debug(calculateDiscreteContinuousUnlabelledMI)
missingTest2 = testData %>% 
        group_by(test) %>% 
        expectFixedSamples(vars(outcome),600)

testData %>% group_by(test) %>% adjustDiscreteMIForAbsentValues(
        vars(outcome), 
        calculateDiscreteContinuousMI, 
        continuousVar = value, 
        sampleCountDf=missingTest2)

testData %>% group_by(test) %>% calculateDiscreteAbsentValuesMI(vars(outcome), sampleCount=600)
testData %>% group_by(test) %>% calculateDiscretePresentValuesMI(vars(outcome), sampleCount=600)

terminological/tidy-info-stats documentation built on Nov. 19, 2022, 11:23 p.m.