问题描述
我正在尝试使用轮廓法来获取我的最佳簇数。但是我收到以下错误：Error in sil.obj[,1:3] : incorrect number of dimensions
这是一个伪数据集：
structure(list(co = c("a","d1","d2","he","hy","ki","li","lu","ob"),c_p = c(6.60985975350616,1.35161744022503,3.56536502546689,4.3010752688172,4.7482646922721,9,8.89473684210526,8.26600985221675,4.93449595916631),ch = c(2.2694432639184,1.37552742616034,5.9592529711375,1.505376344086,1.4173993521518,2.8,2.6917293233083,1.4729064039409,9.0238196512122),co = c(2.7964300892478,1.6033755274262,4.5195246179966,0.3225806451613,2.8505321610366,4.4,8.4962406015038,2.7093596059113,4.6201616333475),di = c(2.311942201445,0.65682137834037,1.062818336163,9.8924731182796,2.1064322073114,2.4,1.0451127819549,8.7192118226601,2.5861335601872),fa = c(5.7161070973226,10.478199718706,46.3497453310696,50.8064516129032,47.5474317445627,51.6,51.1278195488722,50.4926108374384,4.0646533390047),he = c(5.7437314067148,8.64978902953587,8.7096774193548,4.1182795698925,3.1744562702453,48,4.3684210526316,4.6108374384236,4.9472564866014),its = c(13.4296642583935,1.40646976090014,9.08319185059423,1.7096774193548,12.1008792225821,1.4,1.781954887218,10.3448275862069,1.8553806890685),jo = c(6.1198470038249,0.843881856540084,4.15959252971138,3.76344086021505,5.25219805645535,5.2,8.27067669172932,5.78817733990148,4.26414291790727),loss_a = c(7.3523161920952,1.75808720112518,6.7911714770798,8.06451612903226,6.87181860249884,8,9.02255639097744,5.29556650246305,7.60314759676733),loss_s = c(27.2843178920527,5.62587904360056,29.6264855687606,30.1075268817204,2.8857010643221,30.8,25.187969924812,2.6453201970443,3.9642705231816),mu = c(40.3739906502337,7.31364275668073,3.46519524618,41.1290322580645,3.1124479407682,46.4,5.4887218045113,32.7586206896552,3.8987664823479),na = c(39.8215044623884,5.76652601969058,29.7113752122241,3.9032258064516,3.9888940305414,39.2,3.203007518797,36.6995073891626,4.0893236920459),nau = c(6.50233744156396,2.0393811533052,5.51782682512733,6.45161290322581,4.90513651087459,6.01503759398496,6.2807881773399,4.96597192683964),shor = c(31.4067148321292,3.44585091420534,1.9405772495756,18.5483870967742,16.658954187876,20.3007518796992,31.1576354679803,16.5142492556359),sor = c(44.4963875903102,7.73558368495077,6.3327674023769,43.5483870967742,41.8093475242943,41.2,46.6165413533835,3.4088669950739,43.1412165036155),spu = c(22.3969400764981,3.65682137834037,17.062818336163,1.010752688172,16.3350300786673,17.6,1.4210526315789,2.0935960591133,16.3866439812846),temp = c(25.116872078198,9.28270042194093,37.6910016977929,24.7311827956989,2.0245256825544,3.6,3.4511278195489,25.9852216748768,26.1697150148873)),row.names = c(NA,-9L),groups = structure(list(com = c("a",.rows = structure(list(1L,2L,3L,4L,5L,6L,7L,8L,9L),ptype = integer(0),class = c("vctrs_list_of","vctrs_vctr","list"))),class = c("tbl_df","tbl","data.frame"),.drop = TRUE),class = c("grouped_df","tbl_df","data.frame"))
这是我正在运行的代码：
fviz_nbclust(dt_scaled,FUN = hcut,method = "silhouette")
这是我正在使用的图书馆：
library(NbClust)
有人试图用R中的轮廓法获得最佳数目的聚类时遇到相同的东西吗？
解决方法
问题是k.max中fviz_nbclust的默认值为10，这比您拥有的数据（9行）多一组。同样，第一个变量是字符而不是数字。试试这个
fviz_nbclust(dt_scaled,FUN = hcut,method = "silhouette",k.max=5)
或任何小于您数据中观察值的k.max值。