README.md

In BouzidiImen/ClusteringR: Clustering package

ClusteringR

Its functionality includes exploratory data analysis, data segmentation and data visualization. It is designed to handle realistic data sets : hedonic data set and sensory data set. It makes use of several clustering methods as well as the implementation of partition-validity approach.

Finally, a graphical user interface is implemented with R shiny in order to propose a user friendly package.

Installation

You can install the development version from GitHub with :

install.packages("devtools")
devtools::install_github("BouzidiImen/ClusteringR")

Usage

You can find below an overall look at how ClusteringR can be useful for your sensory analysis.

Clustering :

Diverse methods of clustering are available in the ClusteringR package :

'hierarchical', 'diana', 'kmeans', 'clara', 'pam', 'sota' and 'som'

library(ClusteringR)
# Create a clustering object  -------------------------------------------------
cl <- Clustering(t(hedo),ClustMeth='hierarchical',k=3,Hdismethod='euclidean',Hmethod="ward.D2",
                    Graph=F,VarCart=F,IndCart=F )
# get clusters
clusters=cl$classes

#Plot of  dendrogram 

plot(cl$dendrogram)

help("Clustering") # to see more information about the function of clustering 

Sensory Analysis :

Based on the sensory map, this package make it easier to know consumers's behaviour, their likes and dislikes.

library(ClusteringR)
# Create an EPM object  -------------------------------------------------
E <- EPM(hedo,senso,ModelType='Quadratic',respt=FALSE,nbpoints=50,Graphpred=FALSE,Graph2D=TRUE,Graph3D=FALSE,statistic.Value.Type='rsquared')

help("EPM") # to see more information about the function of external preferences mapping 

Data available in the package

• senso : sensory data from a professional trial of 8 biscuits by 12 judges.
• hedo : hedonic data from a trial of 8 biscuits by 294 customers.

#Usage 
library(ClusteringR)
S=senso # sensory data 
H=hedo # hedonic data

A User Friendly Package

Within the package you find a shiny application that demonstrate what the package does and make its use easier.

ClustShiny() #run shiny application

P.S : You can visit the following link to get a sneak peek on the package functionalities.

Shiny application for the package

Acknowledgement :

In preparation of my package, I had to take the help and guidance of my professor Ibtihel Rebhi, who deserves my deepest gratitude.

BouzidiImen/ClusteringR documentation built on March 22, 2022, 8:50 p.m.