Scribd Logo
Upload

Welcome to Scribd!

  • 52 views

    R

    Uploaded by

    Shrutika Agrawal
    This document demonstrates various time series analysis and forecasting techniques using R packages like forecast, fpp2, and ggplot2. It includes examples of plotting, decomposing, smoothing, and modeling time series data like AirPassengers, qcement, arrivals, mink, USAccDeaths, savings.ts, goog, uschange, ausbeer, and mens400. It also covers topics like exponential smoothing, time series linear regression, cross validation, and forecasting. Finally, it discusses clustering, classification, and association rule mining techniques using packages like arules, rattle, ipred, and e1071.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd

    R

    Uploaded by

    Shrutika Agrawal
    52 views33 pages
    This document demonstrates various time series analysis and forecasting techniques using R packages like forecast, fpp2, and ggplot2. It includes examples of plotting, decomposing, smoothing, and modeling time series data like AirPassengers, qcement, arrivals, mink, USAccDeaths, savings.ts, goog, uschange, ausbeer, and mens400. It also covers topics like exponential smoothing, time series linear regression, cross validation, and forecasting. Finally, it discusses clustering, classification, and association rule mining techniques using packages like arules, rattle, ipred, and e1071.

    Copyright

    © © All Rights Reserved

    Available Formats

    DOCX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document demonstrates various time series analysis and forecasting techniques using R packages like forecast, fpp2, and ggplot2. It includes examples of plotting, decomposing, smoothing, and modeling time series data like AirPassengers, qcement, arrivals, mink, USAccDeaths, savings.ts, goog, uschange, ausbeer, and mens400. It also covers topics like exponential smoothing, time series linear regression, cross validation, and forecasting. Finally, it discusses clustering, classification, and association rule mining techniques using packages like arules, rattle, ipred, and e1071.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    52 views33 pages

    R

    Uploaded by

    Shrutika Agrawal
    This document demonstrates various time series analysis and forecasting techniques using R packages like forecast, fpp2, and ggplot2. It includes examples of plotting, decomposing, smoothing, and modeling time series data like AirPassengers, qcement, arrivals, mink, USAccDeaths, savings.ts, goog, uschange, ausbeer, and mens400. It also covers topics like exponential smoothing, time series linear regression, cross validation, and forecasting. Finally, it discusses clustering, classification, and association rule mining techniques using packages like arules, rattle, ipred, and e1071.

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    of 33

    tseries <- c(2,4,6,8,10,12,14,16,18,20)

    barplot(tseries)

    install.packages("forecast")

    install.pacakages("ggplot2")

    install.packages("fpp2")

    AirPassengers

    library(ggfortify)

    ggplot2::autoplot(AirPassengers)

    library(fpp2)

    qcement

    autoplot(qcement)

    ggseasonplot(qcement)

    ggseasonplot(qcement,polar=TRUE)

    ggsubseriesplot(qcement)

    gglagplot(qcement)

    ggAcf(qcement)

    Acf(qcement)

    arrivals

    autoplot(arrivals)

    autoplot(arrivals, facets =TRUE)

    autoplot(arrivals, facets =TRUE) +geom_smooth()

    autoplot(arrivals, facets =TRUE) +geom_smooth() + labs(y='Arrivals (in thousands)')

    summary(arrivals)
    ggAcf(arrivals)

    Acf(arrivals)

    mink

    ggAcf(mink)

    Acf(mink)

    USAccDeaths

    ggAcf(USAccDeaths)

    Acf(USAccDeaths)

    wn <- ts(rnorm(36))

    autoplot(wn)

    acf(wn)

    amount <- c(9,8,9,12,9,12,11,7,13,9,11,10)

    summary(amount)

    #moving average

    economics

    savings.ts <- ts(economics$psavert, start= c(1967,7), frequency=12)

    autoplot(savings.ts)

    autoplot(savings.ts, series='Original') +
    autolayer(ma(savings.ts,3), series='MA3')+

    autolayer(ma(savings.ts,4), series='MA4')+

    xlab('Date')+ ylab('Savings Rate')

    savings.ts

    #EXPONENTIAL SMOOTHING METHOD

    ses(goog)

    ses(goog, alpha=.2, h=1)

    autoplot(ses(goog,alpha=0.2,h=1))

    autoplot(ses(goog,alpha=0.2,h=10))

    uschange

    autoplot(uschange)

    autoplot(uschange[,c('Consumption','Income')])

    autoplot(uschange[,c('Consumption','Income')]) +ylab('%change')

    uschange%>%

    as.data.frame()%>%

    ggplot(aes(x=Income,y=Consumption))+

    ylab('Consumption')+

    xlab('Income')+

    geom_point()
    uschange%>%

    as.data.frame()%>%

    ggplot(aes(x=Income,y=Consumption))+

    ylab('Consumption')+

    xlab('Income')+

    geom_point()+

    geom_smooth()

    uschange%>%

    as.data.frame()%>%

    ggplot(aes(x=Income,y=Consumption))+

    ylab('Consumption')+

    xlab('Income')+

    geom_point()+

    geom_smooth(method='lm', se=FALSE)

    #TIME SERIES LINEAR REGRESSION

    tslm(Consumption~ Income,data=uschange) #consumption=0.5451+0.2806*X

    ##income and consumption are directly proportional

    head(uschange)

    tslm(Consumption ~ ., data=uschange)
    #AUS BEER DATASET

    ausbeer

    autoplot(ausbeer)

    seasonplot(ausbeer)

    beer1992 <- window(ausbeer,start=1992)

    autoplot(beer1992)

    beer.model <- tslm(beer1992 ~ trend+season)

    summary(beer.model)

    autoplot(beer1992, series='Data')+

    autolayer(fitted(beer.model),series='Fitted')+

    xlab('Year')+

    ylab('Megaliters')

    autoplot(beer1992, series='Data')+

    autolayer(fitted(beer.model, n.head=10),series='Fitted')+

    xlab('Year')+

    ylab('Megaliters')

    head(uschange)

    #create a TSLM model to explain the variation in consumption based on all other predictors
    > consumption.model <-
    tslm(Consumption~Income+Production+Savings+Unemployment,data=uschang
    e)

    > consumption.model
    Fitted(consumption.model)

    Summary(consumption.model)

    CV – cross validation

    > ausbeer
    beer1992 <- window(ausbeer,start = 1992)

    autoplot(beer1992)
    beer.model <- tslm(beer1992 ~ trend + season)
    summary(beer.model)
    autoplot(beer1992, series = "Data") + autolayer(fitted(beer.model), series = "Fitted")+
    xlab("Year")+ ylab("Megaliters")
    fitted(beer.model)
    CV(beer.model
    Creating a linear model for marathon dataset

    mens400

    #MENS400 dataset:

    mens400

    autoplot(mens400)

    #to create a mode that explains the rate of changes in winning times

    #forecast the winning time for 2020 and 2024

    mens400.model <- tslm(mens400~ trend)

    mens400.model

    autoplot(mens400)+

    autolayer(fitted(mens400.model))

    forecast(mens400.model)
    A10

    Autoplot(a10)

    Double differenece

    Kwiatkowski-Phillips-Schmidt-Shin (KPSS) test


    1 meaning the series is non stationary

    0 means series is stationary

    Auto regressive model


    Min support =3
    Install.packages(“arules”)

    Install.packages(“lubridate”)

    Intall.packages(“arulesviz”)

    Data(groceries)

    Class(groceries)

    head(groceires,5)

    Fitec <- eclat(groceries,parameter = list(sup=0.1)) (clustering function)

    Fitec1 <- eclat(groceries,parameter = list(sup=0.1,maxlen=15)) (clustering function)

    Inspect(fitec1)
    To check which are highly correlated

    To Explore rules that indicate the sale of soft cheese

    To Explore rules that indicate the sale of yoghurt


    To visualize rules for coffee-

    Rule<-apriori(Groceries,parameter=list(sup=0.001, conf=0.2), appearance = list(rhs='coffee'))

    inspect(head(Rule,5))

    plot(Rule,method="graph",engine='htmlwidget')

    plot(Rule,method="paracoord")

    To identify data based on invoice number-

    To remove the first column invoice number-


    Converting to csv-

    To make data in transaction format-

    DECISION TREE

    install.packages(“rattle”)
    rattle()

    FOR BOOTSTRAP AGGREAGTION

    install.packages(“ipred”)
    Install.packages(“C50”)
    install.packages("e1071")

    library(e1071)

    > library(ISLR)

    > head(Default)
    CLUSTERING

    Install.packages(“tidyverse”)

    Usa<-USArrests

    Summary(usa)

    usa<-scale(usa)

    distances<-get_dist(usa)

    fviz_dist(distances)

    > kc <- kmeans(usa,centers=3,nstart = 20)

    > kc

    fviz_cluster(kc,data=usa)
    > kc3 <- kmeans(usa,centers=3,nstart = 20)

    > kc4 <- kmeans(usa,centers=4,nstart = 20)

    > kc5 <- kmeans(usa,centers=5,nstart = 20)

    > kc6 <- kmeans(usa,centers=6,nstart = 20)

    > wss<-c(kc3$tot.withinss,kc4$tot.withinss,kc5$tot.withinss,kc6$tot.withinss)

    > nclusters<-3:6

    > plot(nclusters,wss,type='b')
    https://cran.r-project.org/web/packages/cluster.datasets/cluster.datasets.pdf
    > usa<-USArrests

    > usa<-scale(USArrests)

    > d<-dist(usa,method = 'euclidean')

    > hcl<-hclust(d,method = "complete")


    > hcl

    > summary(hcl)

    > plot(hcl,cex=0.6,hang=-1)

    For divisive

    To cut the tree


    Tanglegram
    boxplot(mpg~cyl, data=mtcars)
    hist(iris$Sepal.Width,col="blue",main="Histogram",xlab="Sepal width")
    Install.packages(“mvoutlier”)
    Map Tracker plot

    Principal component:
    library(dbscan)

    You might also like