tutorial.Rmd

---
title: "Tutorial"
author: "Rémi Decoster, Pauline Laine, Romain Pic"
date: "08/12/2020"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

## Réalisation d'un clustering k-means 
```{r}
library(ClusterCharPlot)
library(readxl)
library(stats)
```

```{r}
df<- read_excel('Autos.xlsx')
X<-df[,c('length','width','height','engine-size','compression-ratio','horsepower','city-mpg','highway-mpg')]
clus<- kmeans(X,3)
y <-clus$cluster
```

## Caractérisation univarié 

```{r}
#Création de l'objet de la classe UniChar à partir de notre jeu de données complet
obj <- Perform_UniChar(df,y)
plot(obj)
```
```{r}
#Analyse pour les données qualitative
vCramer <- vcramer.UniChar(obj,y)
print(vCramer)
vCramer["label"] <- obj$catnames
colnames(vCramer[1])
barplt.UniChar(obj, vCramer, TRUE)

vtest.UniChar(obj,y, 'make')
#unable to plot yet
```
```{r}
#Analyse pour les données 
#Correlation, caracterisation des partitions
cor <- correlation(obj,y)
print(cor)

#Barplot de la correlation
tmp <- as.data.frame(cor)
(correlations <- t(tmp[4,]))
ClusterCharPlot::barplt.UniChar(obj, correlations)
```


```{r}
#Valeur test, caracterisation des groupes
vt <- valuetest(obj,y)
print(vt)
radar(vt)
```


```{r}
#Effect size, caracterisation des groupes
es <- effectsize(obj,y)
print(es)
radar(es)
```

## Carastérisation multivarié
```{r}
#Création de l'objet multivarié
multivariate=multivariate_charac(X,y)

#Printing and plotting the results of multivariate_charac
print(multivariate$distance_matrix)
print(multivariate$confusion_matrix)
print(multivariate$correlation)
print(multivariate$squared_correlation)
plot(multivariate$correlation)
```

## Calcul des mesures

```{r}
#Création de l'objet
measures=indexes(X,y)

#Printing and plotting the results of indexes
print(measures$internal)
print(measures$external)
plot(measures)
```