80% in Kaggle’s Titanic competition in 50 lines of R code
By Data Tricks, 16 July 2019
Last modified 14 June 2020
Anyone new to machine learning will have probably come across Kaggle’s titanic competition. The task involves applying machine learning techniques to predict which passengers survived the tragedy.
Whilst not a comprehensive attempt to solve the problem, this tutorial guides you through some simple methods to clean the data, engineer features and train an ML algorithm in R to achieve an accuracy of over 80%.
First we need to load the packages required for preparing the data and applying machine learning algorithms. If you don’t have all of the packages use install.packages(“name-of-package”).
For the next part you’ll need to download the Train and Test datasets from Kaggle. Once we’ve loaded the data into R using read.csv, we can bind them with rbind.
train <- read.csv("train.csv")
test <- read.csv("test.csv")
test$Survived <- NA
data <- rbind(train, test)
apply(data, 2, function(x) sum(is.na(x)))
The last line in the code above will print the number of missing values there are in the data, which are as follows:
Survived = 418 (expected as we have included the Test dataset)
Age = 263
Fare = 1
With so many missing values in the Age variable, it is worth applying some technique to impute the values. To keep things simple we’ll just replace all missing values in the Age variable with the mean age for the population, and do the same for Fare.
data$Age[is.na(data$Age)] <- mean(data$Age, na.rm=TRUE)
data$Fare[is.na(data$Fare)] <- mean(data$Fare, na.rm=TRUE)
Tags: kaggle, machine learning, R, titanic
Please note that your first comment on this site will be moderated, after which you will be able to comment freely.