Criar clusters em R

-
Venho fazendo análise de dados con clusters em R e estes são os meus códigos y dataset.

Aqui estou identificando o número de clusters
setwd("/home/sergio/Dropbox/PhD/tesis/flujos1/")
r2 <- read.csv("rotas2.csv", h=T)
names(r2)
attach(r2)
library(NbClust)
library(fpc)
library(cluster)
library(ggplot2)
#Determinando el numero de clusters
#Distance Euclidean
NbClust(pro,diss="NULL", distance="euclidean",min.nc=2,
max.nc=8,method="ward",
index="ch",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="euclidean",min.nc=2,
max.nc=8,method="ward",
index="gamma",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="euclidean",min.nc=2,
max.nc=8,method="ward",
index="gplus",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="euclidean",min.nc=2,
max.nc=8,method="ward",
index="cindex",alphaBeale=0.1)
#8
#Distance maximum
NbClust(pro,diss="NULL", distance="maximum",min.nc=2,
max.nc=8,method="ward",
index="ch",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="maximum",min.nc=2,
max.nc=8,method="ward",
index="gamma",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="maximum",min.nc=2,
max.nc=8,method="ward",
index="gplus",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="maximum",min.nc=2,
max.nc=8,method="ward",
index="cindex",alphaBeale=0.1)
#Distance manhattan
NbClust(pro,diss="NULL", distance="manhattan",min.nc=2,
max.nc=8,method="ward",
index="ch",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="manhattan",min.nc=2,
max.nc=8,method="ward",
index="gamma",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="manhattan",min.nc=2,
max.nc=8,method="ward",
index="gplus",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="manhattan",min.nc=2,
max.nc=8,method="ward",
index="cindex",alphaBeale=0.1)
#Distance Euclidean, method single
NbClust(pro,diss="NULL", distance="euclidean",min.nc=2,
max.nc=8,method="single",
index="ch",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="euclidean",min.nc=2,
max.nc=8,method="single",
index="gamma",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="euclidean",min.nc=2,
max.nc=8,method="single",
index="gplus",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="euclidean",min.nc=2,
max.nc=8,method="single",
index="cindex",alphaBeale=0.1)
#Distance Euclidean, method complete
NbClust(pro,diss="NULL", distance="euclidean",min.nc=2,
max.nc=8,method="complete",
index="ch",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="euclidean",min.nc=2,
max.nc=8,method="complete",
index="gamma",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="euclidean",min.nc=2,
max.nc=8,method="complete",
index="gplus",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="euclidean",min.nc=2,
max.nc=8,method="complete",
index="cindex",alphaBeale=0.1)
#Distance Euclidean, method average
NbClust(med,diss="NULL", distance="euclidean",min.nc=2,
max.nc=8,method="average",
index="ch",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="euclidean",min.nc=2,
max.nc=8,method="average",
index="gamma",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="euclidean",min.nc=2,
max.nc=8,method="average",
index="gplus",alphaBeale=0.1)
NbClust(pro,diss="NULL", distance="euclidean",min.nc=2,
max.nc=8,method="average",
index="cindex",alphaBeale=0.1)
#8 Cluster
view raw gistfile1.r hosted with ❤ by GitHub

8

Então códigos gráficos jerárquiso são:


#Hierarchical Agglomerative
d <- dist(pro, method = "euclidean")
fit <- hclust(d, method="ward")
plot(fit)
rect.hclust(fit, k=8, border="red")
view raw cluster1.r hosted with ❤ by GitHub
E o gráfico é, e eu posso ver, especificamente, como meus dados são classificados

Comentarios

Publicar un comentario

Entradas populares de este blog

ERROR: Grib2 file or date problem, stopping in edition_num. Can't ungrib.exe SOLVED

-
Imagen
Well, it seems that this problem can be very common. Usually, you have a directory for downloading the meteorology from NCEP/UCAR. Probably, you will have the script for downloading the data in the same directory. If you link the grib files based on the name there won't be any problems ./link_grib.csh /path/GFS* However, it could be possible to link the whole directory. You might do this when you have a whole year of meteorological data. ./link_grib.csh /path/ If you have scripts for downloading the met data in the same directory, THEN you will have the message: ERROR: Grib2 file or date problem, stopping in edition_num To avoid this error, just move the script for downloading the met data elsewhere and link the directory again. Finally, do ungrib.exe UPDATE 2018-12/15 (I mean 15 of December of 2018) I had the same problem again, and with the same data! and this solution did not work till i did this: ./link_grib.csh /path/GFS Yes only the first letters, in ...
Leer más

ARCGIS: Cálculo de pendiente entre puntos en el espacio xyz

-
Imagen
Se tienen puntos en el espacio con coordenadas xyz. Estas corresponden a coordenadas UTM huso 19, en metros. Pendiente = Distancia Vertical / Distancia Horizontal  Distancia Vertical = z2 - z1 Distancia Horizontal = d = [(y2-y1)^2 + (x2-x1)^2]^0.5 Pendiente = (z2 - z1) / [(y2-y1)^2 + (x2-x1)^2]^0.5 Pendiente porcentual = Pendiente *100 Entonces; x1 y1 z1 x2 y2 z2 pendiente% 344150.34 6302722 520.897 344911.2 6302394 530.5 1.159 344150.34 6302722 520.897 343774.7 6302378 521.858 0.188 345811.8 6302371 536 345893.9 6302547 533 -1.544 Sirve para calcular pendiente de caminos, ríos, etc :) En este caso lo hice para una modelación de flujo vehicular de Santiago Estraus , por lo que la diferencias de las alturas está según el sentido del flujo, desde la altura del nodo inicial al nodo final, por eso hacia el sector oriente se aprecia mayor pendie...
Leer más

WRF open_aux_u : error opening auxinput5_d02_2013-12-31_00:00:00 for reading. 100

-
Imagen
So, i got this error: open_aux_u Timing for Writing /scr4/sibarra/masp/health/wrfo for domain        2:    4.14512 elapsed seconds open_aux_u : error opening auxinput5_d02_2013-12-31_00:00:00 for reading. 100 pretty weird because it a met run. According to http://forum.wrfforum.com/viewtopic.php?f=47&t=10003, they recommend diminishing the number of processors. I ran with 32 and got stuck I ran with 16 and appear this error open_aux_u : error opening auxinput5_d02_2013-12-31_00:00:00 for reading. 100 I'm running with  8 and still running. 
Leer más