OptClusters: OptClusters
In Sautie/MALDITOFSpectraPA: Processing and analysis of MALDI_TOF spectra

Description Usage Arguments Value Examples

OptClusters is a wrapper of several functions to visualize and compute optimal clusters for different clustering and evaluation methods

OptClusters(
  df_m,
  meth = "kmeans",
  dist = "euclidean",
  varCat1,
  value,
  minc = 2,
  maxc = 10,
  ind = "average silhouette width",
  nb = 100
)

`df_m:`	dataframe containing peaks and metadata
`meth:`	clustering algorithms (meth="kmeans", default value), other values: "pam" or "hclust"
`dist:`	distances ("euclidean", default), "maximum", "manhattan", "canberra", "binary" "minkowski"
`varCat1:`	categorical variable for choosing isolates, examples: "Taxonomie" ,"Genre", "Date.d.analyse" ,"Origine","Ruche", "Nutrition" , "Date.de.récolte" , "Lieu.de.la.ruche"
`value:`	level of catVar1, examples: "Lactobacillus" ("Genre"), Taxonomie("Pediococcus pentosaceus"), "Erica cinerea" ("Nutrition"),...
`minc:`	minimal number of clusters (minc=2, default value)
`maxc:`	maximal number of clusters (maxc=10, default value)
`ind:`	methods to evaluate clustering algorithms:"total within sum of squares" ,"average silhouette width" and "gap statistics"
`nb:`	number of bootstrap samples (nb=100, default value)

figures and statistics

     OptClusters(df_Peaks, varCat1="Taxonomie"	, value="Enterococcus faecalis"),
               OptClusters(df_Peaks, varCat1="Taxonomie"	, value="All")
               OptClusters(df_Peaks, meth="pam", varCat1="Taxonomie"	, value="All", ind="gap statistics"),
               OptClusters(df_Peaks, meth="hclust", varCat1="Taxonomie"	, value="All", ind="gap statistics")
       source: https://rpubs.com/pg2000in/OptimumClusters
               http://rstudio-pubs-static.s3.amazonaws.com/265632_3ad9e0b981244e15887677f8dffb39a0.html#using-30-different-indices
               https://www.datanovia.com/en/lessons/determining-the-optimal-number-of-clusters-3-must-know-methods/
               https://www.rdocumentation.org/packages/factoextra/versions/1.0.7/topics/fviz_nbclust