Clasificación

library(dplyr)
library(ggplot2)
library(Amelia)

train<-read.csv("train.csv",stringsAsFactors = FALSE, na.strings = "")
test<-read.csv("test.csv",stringsAsFactors = FALSE, na.strings = "")

# Agrega como SouthHampton a Embark si falta la data
train$Embarked<-ifelse(is.na(train$Embarked),
                       yes = "S",
                       no = train$Embarked)
test$Embarked<-ifelse(is.na(test$Embarked),
                      yes = "S",
                      no = test$Embarked)
missmap(train)

# Función para extraer el titulo del nombre
extract_title <- function(name){
  y<-strsplit(name,split = "[.,]")
  y<-y[[1]][2]
  return(trimws(y))
}

# Agregamos la columna del titulo
train$Title<-sapply(train$Name, FUN = extract_title,USE.NAMES = FALSE)
test$Title<-sapply(test$Name, FUN = extract_title,USE.NAMES = FALSE)

table(train$Title, train$Sex)

train$Title<-ifelse(train$Title %in% c("Miss","Mlle","Lady","Ms"),
                    yes = "Miss",
                    no =  train$Title)

test$Title<-ifelse(test$Title %in% c("Miss","Mlle", "Lady", "Ms"),
                   yes = "Miss",
                   no =  test$Title)

train$Title<-ifelse(train$Title %in% c("Mrs","Mme", "Dona"),
                    yes = "Mrs",
                    no =  train$Title)

test$Title<-ifelse(test$Title %in% c("Mrs","Mme", "Dona"),
                   yes = "Mrs",
                   no =  test$Title)



train$Adult<-ifelse(train$Age > 12,yes = 1, no = 0)
test$Adult<-ifelse(test$Age > 12,yes = 1, no = 0)

train$Adult<-ifelse(is.na(train$Age) & trimws(train$Title) == "Master" ,
                    yes = 0,
                    no = ifelse(is.na(train$Age),
                                yes = 1,
                                no = train$Adult)   
)
test$Adult<-ifelse(is.na(test$Age) & trimws(test$Title) == "Master" ,
                   yes = 0,
                   no = ifelse(is.na(test$Age),
                               yes = 1,
                               no = test$Adult)   
)


summary(train$Fare)

test$Fare <- ifelse(is.na(test$Fare),
                    yes = median(train$Fare),
                    no=test$Fare)

modelo1<-lm(data = train, 
            formula = Survived ~ Adult +
              Title + Embarked + Fare +
              Sex + Pclass)

prediccion<-predict(object = modelo1,newdata = test)
prediccion <- ifelse(prediccion>0.5, yes = 1, no = 0)
prediccion <- data.frame(PassengerId = test$PassengerId,
                          Survived = prediccion)

write.csv(prediccion, file="ModAdultTitle.csv",row.names = FALSE)