Skip to content

Update deep learning model and funtions #138

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
wants to merge 1 commit into
base: OHW21
Choose a base branch
from
Open
Show file tree
Hide file tree
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
Binary file added deep-learning/.RData
Binary file not shown.
385 changes: 385 additions & 0 deletions deep-learning/.Rhistory
Original file line number Diff line number Diff line change
@@ -0,0 +1,385 @@
# Este es un comentario; en R los comentarios empiezan
# a partir del caracter '#'.
# -------------------
# En seguida asignaremos mediante código de R el valor 2014 a una
# variable llamada 'x':
x <- 2014
# Ahora se imprimirá el valor de la variable dos veces, la primera
# vez se hará de manera explícita por medio de la función print(),
# como sigue:
print(x)
## [1] 2014
# ... en seguida se imprimirá de manera implícita, simplemente
# 'tecleándola' en la consola:
x
## [1] 2014
# Finalmente haremos una multiplicación de x por 2
2*x
#AYUDA EN R
help(lm)
# Una forma abreviada sería
?lm # -- comentario
help("[[")
help('if')
help.search("split")
# O abreviadamente:
??"split"
#AYUDA EN BLOGS
help<-stackoverflow-rbloggers
#========NUMEROS REALES=========#
x <- 2 # Se asigna el valor 2 a x
print(x) # Se imprime el valor de x
class(x) # Muestra cuál es la clase de x
x <- 6/2 # Se asigna el valor de la operación dividir 6/2 a x
print(x)
class(x)
#========NUMEROS ENTEROS=========#
x <- as.integer(6/2); print(x) #Convertir la división a entero mediante la funcion
class(x)
#======NUMEROS COMPLEJOS==========#
x <- 21 + 2i
y <- 2i + 21 # El mismo valor que x
z <- -1 + 0i # Corresponde a -1
tt <- sqrt(z) # raíz cuadrada de -1
print(x); print(y); print(z); print(tt)
class(tt)
x <- 1/0 # División por cero
x
## [1] Inf
# Tambien dividir un número por Inf da cero:
y <- 1/Inf
y
## [1] 0
#Finalmente, algunas operaciones pueden resultar en algo que no es un número,
#esto se representa por el valor NaN. Veamos un ejemplo:
x <- 0/0
x
#===============VECTORES========================#
x <- 2 # Se asigna el valor 2 a x
print(x) # Se imprime el valor de x
c(4,2,-8) # Creación de un vector sin asignarlo a una variable
## Distintas formas de asignar un vector a una variable
u <- c(4,2,-8) # Usando el operador <-
c(4, 2, -8) -> v # Usando el operador ->
# Usando la función assign:
assign("w", c(4, 2, -8))
p = c(4, 2, -8) # Usando el operador =
print(u); print(v); print(w); print(p)
#Creación de vectores a partir de archivos de texto - la función scan()
#12 15.5 3.1
#-2.2 0 0.0007
vec <- scan("UnVec.txt")
print(vec)
vec <- scan("IntVec.txt", integer())
print(vec); class(vec) # El vector y su clase
vv <- c(5, 6.6, -7.7)
write(vv, "OtroArchivo.txt")
# Ahora recuperemos el contenido del archivo
v1 <- scan("OtroArchivo.txt")
v1
#========Creación de vectores a partir de secuencias y otros patrones========#
v <- vector("integer", 0)
v # Un vector de enteros sin elementos
w <- vector("numeric", 3)
w # Un vector de tres ceros
u <- vector("logical", 5)
u # Un vector de 5 FALSE
#=========OPERADOR ":" =========#
1:3
v <- 40:13
print(v); class(v) # El vector y su clase
#=========FUNCION seq()========#
#seq() que permite generar una mayor variedad de secuencias numéricas.
v <- seq(from = 5, to = 15, by = 2)
print(v) # secuencia desde 5 hasta 15 de 2 en 2
v <- seq(from = 4, by = 2, length.out = 8)
print(v) # secuencia de 8 números iniciando desde 4 y de 2 en 2
v <- c(4, 8, -3) #Repetir la misma secuencia 5 veces.
w <- rep(v, times = 5)
print(w)
#construir un vector a partir de dos o más vectores ya existentes
u <- c(3, 4, 5)
v <- c(5, 4, 3)
w <- c(u, v)
print(w) # La concatenación de u y v
#Acceso a los elementos individuales de un vector
v <- c(8, 7, -3, 2, 182)
v[5] # El quinto elemento
print(v[1]); print(v[3])
v[4]+v[2] # La suma del cuarto y segundo elementos de v
v[1] <- v[2] - v[5]
v # Note que el resultado de la operación se ha guardado en v[1]
v[8] <- 213
v # v tiene ahora 8 elementos con espacios vacios: NA
#=================================================================#
frutas <- c(15, 100, 2, 30)
frutas
#Supóngase ahora que se quiere asociar esos valores con el nombre de la fruta correspondiente:
names(frutas) <- c("naranja", "pera", "manzana", "durazno")
frutas
frutas <- c(naranja = 15, pera = 100, manzana = 2, durazno = 30)
frutas["durazno"]
frutas["manzana"] <- 8
frutas
frutas[2]
#==============OPERACIONES SENCILLAS CON VECTORES=====================#
v <- 2 + 3 ; v # Resulta en un vector de longitud 1
v <- c(2, 3) - c(5, 1) ;v# Resulta en un vector de longitud 2
v <- c(2, 3, 4) * c(2, 1, 3) ;v # Resulta en un vector de longitud 3
v <- c(2, 3, 4)^(3:1) ;v # Eleva a potencias 3,2,1
#========================FUNCION length()================================#
#En muchas ocasiones es necesario saber la longitud de una vector. La función
#length() aplicada a un vector regresa precisamente ese valor
u <- 2:33 ; v <- c(4, 5, 6) ;w <- c(u, v);w
length(w)
w <= 10 # Prueba elementos menores o iguales a 10
v <- c(4, 5, 6, 7, 8, 9, 10) * c(1, 2)
## Warning: longitud de objeto mayor no es múltiplo de la longitud de uno menor
v
#to be obtained from site
num_fish=12
anchor_files=c('anchors4.csv','anchors5.csv','anchors6.csv','anchors7.csv',
'anchors8.csv','anchors9.csv','anchors10.csv','anchors11.csv','anchors12.csv')
filenames=c('HG_Ecology.csv','HG_Economics.csv','HG_Ethical.csv',
'HG_Institutional.csv','HG_Social.csv','HG_Technology.csv')
nfield=length(filenames)
print("Starting...")
discipline.names =strsplit(filenames, ".csv")
sustainability=matrix(data=0,nrow=num_fish,ncol=nfield)
n_radar=round(num_fish/10,0)
p1_radar=seq(1,n_radar*10,by=10)
p2_radar=p1_radar
p2_radar[1:(n_radar-1)]=p1_radar[2:n_radar]-1
p2_radar[n_radar]=num_fish
for(i in 1:nfield) {
fisheries.all = loaddata(filenames[i])
n_att=ncol(fisheries.all)
anchors=loaddata(anchor_files[n_att-3])
n_an=nrow(anchors)
colnames(anchors)<-colnames(fisheries.all)
fisheries.dat=fisheries.all[1:num_fish,]
fisheries.raw=rbind(anchors,fisheries.dat)
fisheries.scaled = mdscale(fisheries.raw)
f_images=paste(discipline.names,".jpg",sep="")
jpeg(filename=f_images[i],width=20,height=16,units="cm",res=500)
Res=ifelse(nfield>30,
RAPplot1(fisheries.scaled,num_fish,n_an),
RAPplot2(fisheries.scaled,num_fish,n_an))
mtext(side=3, line=1, discipline.names[i],adj=0)
dev.off()
resfiles=paste("Results_",discipline.names,".csv",sep="")
write.csv(fisheries.scaled[(n_an+1):nrow(fisheries.scaled),],resfiles[i])
sustainability[,i]= fisheries.scaled[(n_an+1):nrow(fisheries.scaled),1]
}
rownames(sustainability)<-rownames(fisheries.dat)
colnames(sustainability)<-discipline.names
#Radar plots
for(i in 1:n_radar) {
f_images=paste("Radar_",seq(1,n_radar),".jpg",sep="")
jpeg(filename=f_images[i],width=20,height=16,units="cm",res=500)
radarplot(sustainability[p1_radar[i]:p2_radar[i],])
dev.off()
}
print("Complete.")
View(sustainability)
View(discipline.names)
#####33###
library(FactoMineR)
temperature <- read.table("http://factominer.free.fr/bookV2/temperature.csv",
header=TRUE,sep=";",dec=".",row.names=1)
source("D:/PROGRAMMING/PROGRAMMING WITH R/Temperature.R")
dat<-data.frame(t=seq(0,2*pi,by=0.1))
xcor<-function(t) 16*sin(t)^3
ycor<-function(t) 13*cos(t)-5*cos(2*t)-2*cos(3*t)-cos(4*t)
dat$y=ycor(dat$t)
dat$x=xcor(dat$t)
with(dat,plot(x,y,type="l"))
source("~/.active-rstudio-document")
source("D:/PROGRAMMING/PROGRAMMING WITH R/Temperature.R")
dat<-data.frame(t=seq(0,2*pi,by=0.1))
xcor<-function(t) 16*sin(t)^3
ycor<-function(t) 13*cos(t)-5*cos(2*t)-2*cos(3*t)-cos(4*t)
dat$y=ycor(dat$t)
dat$x=xcor(dat$t)
with(dat,plot(x,y,type="l"))
with(dat,polygon(x,y,col="red",lwd=3,lty = 3))
points(c(10,-10,-15,15),c(-10,-10,10,10),pch=169,font=5,cex=5,col="purple")
text(0,0,"TE QUIERO",col='white',cex = 2.5)
#####33###
library(FactoMineR)
temperature <- read.table("http://factominer.free.fr/bookV2/temperature.csv",
header=TRUE,sep=";",dec=".",row.names=1)
res.pca <- PCA(temperature[1:23,],scale.unit=TRUE,ncp=Inf,
graph=FALSE,quanti.sup=13:16,quali.sup=17)
##############################
ctl <- c(4.17,5.58,5.18,6.11,4.50,4.61,5.17,4.53,5.33,5.14)
trt <- c(4.81,4.17,4.41,3.59,5.87,3.83,6.03,4.89,4.32,4.69)
group <- gl(2, 10, 20, labels = c("Ctl","Trt"))
weight <- c(ctl, trt)
lm.D9 <- lm(weight ~ group)
lm.D90 <- lm(weight ~ group - 1) # omitting intercept
anova(lm.D9)
summary(lm.D90)
opar <- par(mfrow = c(2,2), oma = c(0, 0, 1.1, 0))
plot(lm.D9, las = 1) # Residuals, Fitted, ...
par(opar)
######
counts <- c(18,17,15,20,10,20,25,13,12)
outcome <- gl(3,1,9)
set.seed(123)
set.seed(123)
x <- rnorm(100, 10, 2)
y <- 0.78*x + rnorm(100, 0, 1)
# Crear un data frame con los datos
datos <- data.frame(x, y)
datos
# Ajustar el modelo de regresión lineal simple
modelo <- lm(y ~ x, data = datos)
summary(modelo)
library(tensorflow)
library(keras)
# Cargar el dataset MNIST
mnist <- dataset_mnist()
a <- to_categorical(c(0, 1, 2, 3), num_classes=4)
View(mnist)
??to_categorical
a <- keras::to_categorical(c(0, 1, 2, 3), num_classes=4)
a <- keras3::to_categorical(c(0, 1, 2, 3), num_classes=4)
a
library(keras3)
library(tensorflow)
library(keras)
library(dplyr)
library(keras3)
# Cargar la funciones auxiliares
source("tutorial_functions.R")
setwd("C:\Users\EQUIPO\Documents\GitHub\ohw-tutorials\deep-learning")
setwd("C:/Users/EQUIPO/Documents/GitHub/ohw-tutorials/deep-learning")
library(tensorflow)
library(keras)
library(dplyr)
library(keras3)
# Cargar la funciones auxiliares
source("tutorial_functions.R")
# Leer archivo csv
raw_data <- readr::read_csv("tutorial_data_test.csv")
head(raw_data)
View(raw_data)
# Transformación logaritmica
raw_data <- raw_data %>%
log_inputs(vars = c("t1", "t2", "t3", "t4", "t5", "t6", "t7", "t8", "t9", "t10", "t11", "t12"))
#Generate images from data
image_list <- make_image_list(raw_data,
tox_levels = c(0,10,30,80),
forecast_steps = 1,
n_steps = 2,
minimum_gap = 4,
maximum_gap = 10,
toxins = c("t1", "t2", "t3", "t4", "t5", "t6", "t7", "t8", "t9", "t10", "t11", "t12"),
environmentals = c("sst_cum"))
#Splits image_list by year for grouping into train/test data
years <- sapply(image_list, function(x) {return(x$year)})
#image_list <- split(image_list,as.factor(years))
image_list <- split(image_list, years)
#configuration
YEARS_TRAINING <- c("2014", "2016", "2017")
YEARS_TESTING <- "2015"
a <- keras3::to_categorical(c(0, 1, 2, 3), num_classes=4)
print(a)
# Cargar la funciones auxiliares
source("tutorial_functions.R")
#Make a training set
train <- pool_images_and_labels(image_list[YEARS_TRAINING], num_classes = 4)
#Make a test set
test <- pool_images_and_labels(image_list[YEARS_TESTING], num_classes = 4)
#Make a test set
test <- pool_images_and_labels(image_list[YEARS_TESTING], num_classes = 4)
model <- keras::keras_model_sequential() %>%
keras::layer_dense(units=64,
activation = "relu",
input_shape = dim(train$image)[2],
name = "input_layer") %>%
keras::layer_dropout(rate = 0.4,
name = "dropout_1") %>%
keras::layer_dense(units=32,
activation = "relu",
name = "hidden_1") %>%
keras::layer_dropout(rate=0.3,
name = "dropout_2") %>%
keras::layer_dense(units=16,
activation = "relu",
name = "hidden_2") %>%
keras::layer_dropout(rate=0.2,
name = "dropout_3") %>%
keras::layer_dense(units = 4,
activation = "softmax",
name = "output")
reticulate::py_last_error()
reticulate::py_last_error()$r_trace$full_call
#-----
model <- keras3::keras_model_sequential() %>%
keras3::layer_dense(units=64,
activation = "relu",
input_shape = dim(train$image)[2],
name = "input_layer") %>%
keras3::layer_dropout(rate = 0.4,
name = "dropout_1") %>%
keras3::layer_dense(units=32,
activation = "relu",
name = "hidden_1") %>%
keras3::layer_dropout(rate=0.3,
name = "dropout_2") %>%
keras3::layer_dense(units=16,
activation = "relu",
name = "hidden_2") %>%
keras3::layer_dropout(rate=0.2,
name = "dropout_3") %>%
keras3::layer_dense(units = 4,
activation = "softmax",
name = "output")
summary(model)
str(train)
head(train$labels)
model %>% keras::compile(optimizer = "adam",
loss = "categorical_crossentropy",
metrics = "categorical_accuracy")
history <- model %>%
keras::fit(x = train$image,
y = train$labels,
batch_size = 128,
epochs = 32,
verbose=1,
validation_split = 0.2,
shuffle = TRUE)
history <- model %>%
keras3::fit(x = train$image,
y = train$labels,
batch_size = 128,
epochs = 32,
verbose=1,
validation_split = 0.2,
shuffle = TRUE)
plot(history)
plot(history)
history <- model %>%
keras3::fit(x = train$image,
y = train$labels,
batch_size = 128,
epochs = 32,
verbose=1,
validation_split = 0.2,
shuffle = TRUE)
reticulate::py_last_error()
plot(history)
model %>% keras3::compile(optimizer = "adam",
loss = "categorical_crossentropy",
metrics = "categorical_accuracy")
history <- model %>%
keras3::fit(x = train$image,
y = train$labels,
batch_size = 128,
epochs = 32,
verbose=1,
validation_split = 0.2,
shuffle = TRUE)
plot(history)
Loading