Adding data files and R scripts

Rscripts/eda.R

@@ -0,0 +1,27 @@
+dataset.test1 = read.csv('train_subject1_psd01.csv')
+plot(density(dataset.test1[dataset.test1$X7.00E.00==7,1]))
+lines(density(dataset.test1[dataset.test1$X7.00E.00==2,1]))
+lines(density(dataset.test1[dataset.test1$X7.00E.00==3,1]))
+#lines(density(dataset.test1[,20]))
+
+plot(density(dataset.test1[dataset.test1$X7.00E.00==7,50]))
+lines(density(dataset.test1[dataset.test1$X7.00E.00==2,50]))
+lines(density(dataset.test1[dataset.test1$X7.00E.00==3,50]))
+
+plot(density(dataset.test1[dataset.test1$X7.00E.00==7,72]))
+lines(density(dataset.test1[dataset.test1$X7.00E.00==2,72]))
+lines(density(dataset.test1[dataset.test1$X7.00E.00==3,72]))
+
+correlationMatrix = cor(dataset.test1[,-97])
+correlationDataframe = data.frame(correlationMatrix)
+
+for(i in seq(1,96,1)){
+  for(j in seq(i,96,1)){
+    if(abs(correlationMatrix[i,j])>0.6 && i!=j){
+      print(paste(i,j,correlationMatrix[i,j]))
+    }
+  }
+}
+
+var(dataset.test1)
+apply(dataset.test1,2,var)

data/.Rapp.history

@@ -0,0 +1,263 @@
+q
+c*q2
+p = c*q2
+p + (1.96*q)
+p - (1.96*q)
+p + (1.96*q)
+states <- as.data.frame(#
+  state.x77[,c("Murder","Population",#
+               "Illiteracy", "Income", "Frost")])
+states
+state
+state.x77
+dim(states)
+t(states[1,])
+t(states[1,])#
+dtrain <- states[1:25,]#
+dtest <- states[26:50,]
+murderModel <- lm (Murder ~ Population + Illiteracy #
+             + Income + Frost, data=dtrain)
+summary (murderModel)
+gvlma
+import gvlma
+library(gvlma)
+install.packages('gvlma')
+libarary(gvlma)
+library(gvlma)
+gvlma(lm(Murder ~ Population + Illiteracy #
+             + Income + Frost), data)
+data
+states <- as.data.frame(#
+  state.x77[,c("Murder","Population",#
+               "Illiteracy", "Income", "Frost")])#
+dim(states)#
+t(states[1,])#
+dtrain <- states[1:25,]#
+dtest <- states[26:50,]
+gvlma(lm(Murder ~ Population + Illiteracy #
+             + Income + Frost), dtrain)
+dtrain
+gvlma(lm(Murder ~ Population + Illiteracy + Income + Frost), dtrain)
+gvlma((Murder ~ Population + Illiteracy + Income + Frost), dtrain)
+install.packages('car')
+library(car)
+crPlots
+crPlots(murderModel)
+summary(murderModel)
+qqplot(dtrain)
+qqplot(dtrain['Murder'])
+dtrain['Murder']
+qqplot(dtrain['Population','Illiteracy','Income','Frost'], dtrain['Murder'])
+train['Population','Illiteracy','Income','Frost']
+dtrain['Population','Illiteracy','Income','Frost']
+dtrain['Population','Illiteracy','Income']
+dtrain['Population']
+dtrain['Population','Illiteracy']
+dtrain['Population. Illiteracy']
+dtrain['Population, Illiteracy']
+dtrain['Population + Illiteracy']
+dtrain[c('Population','Illiteracy')]
+qqplot(dtrain[c('Population','Illiteracy','Income','Frost')],dtrain['Murder'])
+qqplot(dtrain['Income'], dtrain['Murder'])
+dtrain['Income']
+gvlma((Murder ~ Population + Illiteracy + Income + Frost), dtrain)
+ans = gvlma((Murder ~ Population + Illiteracy + Income + Frost), dtrain)
+summary(ans)
+plot(ans)
+summary(deletion.gvlma(ans))
+plot(ans)
+summary(ans)
+durbinWatsonTest(murderModel)
+crPlots(murderModel)
+summary(ans)
+vif(murderModel)
+vif(dtrain)
+vif(murderModel)
+outlierTest(murderModel)
+library(MASS)
+chisq.test(dtrain)
+chisq.test(dtrain['Muder'])
+chisq.test(dtrain['Murder'])
+chisq.test(dtrain['Population'])
+??chisq.test
+?chisq.test
+2^2
+2**2
+2**3
+a = [95,85,90]
+a = c(95,85,90)
+a
+mean(a)
+sd(a)
+pnorm(95,mean(a), sd(a))
+exp(-0.5)/(sqrt(2*pi)*5)
+var(a)
+dnorm(95,mean(a), sd(a))
+b = c(125, 100,70,120,60,220,75)
+dnorm(95, mean(a), sd(a))
+dnorm(95, mean(b), sd(b))
+P2 = dnorm(95, mean(b), sd(b))
+P1 = dnorm(95, mean(a), sd(a))
+(P1*3)/(P2*7)
+c(c(2,10), c(2,5), c(8,4), c(5,8), c(7,5), c(6,4), c(1,2), c(4,9))
+t = c(c(2,10), c(2,5), c(8,4), c(5,8), c(7,5), c(6,4), c(1,2), c(4,9))
+t
+cast(t)
+matrix(t, ncol=2)
+matrix(t, ncol=2, byrow=TRUE)
+d = matrix(t, ncol=2, byrow=TRUE)
+d
+plot(d)
+points(2,10, color='red')
+points(2,10)
+points(2,10, col='red')
+points(5,8, col='bluw')
+points(5,8, col='blue')
+points(1,2, col='green')
+d[1]
+d[1,]
+d[c(3,4,5,6,8),]
+mean(d[c(3,4,5,6,8),])
+rowmean(d[c(3,4,5,6,8),])
+rowmeans(d[c(3,4,5,6,8),])
+help(means)
+??means
+colMeans(d[c(3,4,5,6,8),])
+colMeans(d[c(1),])
+colMeans(d[c(1,),])
+d[1, ]
+c(1)
+d[c(1), ]
+colMeans(d[c(1), ])
+colMeans(d[c(2,7), ])
+colMeans(d[c(1,8), ])
+colMeans(d[c(3,4,5,6), ])
+colMeans(d[c(2,7), ])
+kmeans(d)
+kmeans(d, 3)
+kmeans(d, ceners=d[c(1,4,7), ])
+kmeans(d, centers=d[c(1,4,7), ])
+colMeans(d[c(1,4,8), ])
+colMeans(d[c(3,5,6), ])
+c = c(1,2,2,3,5,1,1,3,4,3)
+c
+matrix(c, ncol=2)
+d = matrix(c, ncol=2)
+d[1,]
+d[2,]
+d[1,] - d[2,]
+abs(d[1,] - d[2,])
+max(abs(d[1,] - d[2,]))
+maxnorm <- function(p1, p2) {return (max(abs(p1 - p2)))}
+p = matrix(NA, nrow=5, ncol=2)
+p
+p = matrix(NA, nrow=5, ncol=5)
+for(i in 1:5) {
+for (j in 1:5){
+p[i,j] = maxnorm(d[i,], d[j,])
+}
+p
+hclust(d)
+hclust(mtcars)
+hclust(distr(mtcars))
+hclust(dist(mtcars))
+plot(hclust(dist(mtcars)))
+d1 = rnorm(1000,mean=6, sd=1.5)
+d1
+plot(d1)
+plot(1:1000, d1)
+plot(density(d1))
+d2 = rnorm(n=1000, )
+d1 = rnorm(n=20, mean=6, sd=1.5)
+d2 = rnorm(n=20, mean=4, sd=0.8)
+plot(density(d1), col='red')
+plot(density(d2), col='blue')
+plot(density(d1), col='red')
+par(new=true)
+par(new=TRUE)
+plot(density(d2), col='blue')
+density(d1)
+density(d1)*density(d2)
+density(d1)
+plot(density(d1), col='red')
+lines(density(d2), col='blue')
+type(density(d1))
+typeof(density(d1))
+help(dnorm)
+dnorm(mean=0, sd = 1)
+dnorm(20, mean=0, sd = 1)
+curve(d1)
+curve(density(d1))
+curve(rnorm(20, 6, 1.5))
+density(d1)[0]
+density(d1)[1]
+density(d1)[2]
+density(d1)[3]
+density(d1)[4]
+density(d1)[5]
+density(d1)[6]
+density(d1)$data
+density(d1)$data.name
+plot(density(d1)[1])
+plot(density(d1)[1], density(d1)[2])
+density(d1)
+density(d1)[2]
+plot(1:512, density(d1)[1])
+plot(1:512, density(d1)[2])
+plot(0:512, density(d1)[1])
+dim(512, density(d1)[1])
+dim(density(d1)[1])
+length(density(d1)[1])
+(density(d1)[1])
+(density(d1)[2])
+plot(density(d1)[1]$x, density(d1)[2]$y)
+density(d2)[1]$x
+density(d1)[1]$x
+density(d2)[1]$x
+density(d1)[1]$x == density(d2)[1]$x
+sum(density(d1)[1]$x == density(d2)[1]$x)
+obs <- c(0.4, 0.5, 0.8, 0.1)
+pri <- rnorm(10000, 0, 1)
+likfun <- function(theta) {
+sapply( theta, function(t) prod( dnorm(obs, t, 0.5) ) )
+}
+tmp <- likfun(pri)
+tmp
+n = 20
+mean.pri = 4
+mean.likli = 6
+sd.pri = 0.8
+sd.likli = 1.5
+d1 = rnorm(n, mean.pri, sd.pri)
+tmp = dnorm(d1, mean.likli, sd.likli)
+tmp
+post = sample(pri, 20, prob = tmp)
+length(pri)
+post = sample(d1, 20, prob = tmp)
+post
+plot(density(post))
+density(post)
+mean(post)
+sd(post)
+sigma = (n/(sd.likli^2)) + (1/(sd.pri)^2)
+sigma = 1/sigma
+sigma
+sqrt(sigma)
+d1 = rnorm(10000, mean.pri, sd.pri)
+tmp = dnorm(d1, mean.likli, sd.likli)
+tmp
+plot(temp)
+plot(tmp)
+plot(density(tmp))
+post = sample(d1, 20, prob = tmp)
+plot(density(post))
+var(post)
+sd(post)
+post = sample(d1,5000, prob = tmp)
+sd(post)
+setwd('Google Drive/Courses/Sem 2/CSC 591 - ML/Project/data/')
+ls
+ls()
+d1 = read.csv(file='test_subject1_psd04.csv')
+summary(d1)
+names(d1)