Predicting the Survival of Titanic Passengers

Predict passengers being dead/alive using titanic datasets applying R language and Random forest

This programs explains how to use the titanic datasets to predicts if a passenger is alive or dead. The variables are- survival(0 = No, 1 = yes)

pclass (1 = 1st, 2 = 2nd, 3 = 3rd)

age, sex, sibsp (siblings / spouses aboard the Titanic)

parch (parents / children aboard the Titanic)

ticket numer, fare (passenger fare), cabin (cabin number)

emabraked (C = Cherbourg, Q = Queenstown, S = Southampton)

The steps used to train model on window (10) in the R studio Desktop (R-Studio 8.12 build 175481) envionment are decribed below.

Steps

1. Install R studio Desktop

Visit https://rstudio.com/products/rstudio/ Download and install R studio Desktop

Download Titanic data set from Kaggle

Visit https://www.kaggle.com/c/titanic/data Download train.csv and test.csv Place the downloaded files in a folder (say titanic) in any drive

2. Set up Working Directory and R studio Environment

load the training and testing data sets in to the R studio

3. Data cleaning and Analysis

Loading the train and test data

train <- read.csv("train.csv", header = TRUE)

test <- read.csv("test.csv", header = TRUE)

3a. pclass and survival

3b. Relationship between pclass, survival, and title

If a person title is "Mr" and he happens to be in the third class, he is most likely to perish

The passenger with title "Mr" is more likely to perish wherther he is in 1 st, 2nd, and 3rd class

Most of the females are in the 1st and 2nd class.

Females in the 1st and 2nd class are more likely to survive

3c. Relationship between pclass, sex, and survival

3d Survival rates as sex, pclass, and age

conclusion:

pclass, title, sibsp, parch, and family size varaibles have influence on the passenger survival ticket number, fares, cabin, embarked variables have no influence on the passenger survival pclass, title, sibsp, parch, and family size is used for building machine learning model

4. Building Machine Learning Models

4a. Train a Random Forest (using pclass and title)

4b Train a Random Forest using pclass, title, & family.size

OOB estimate of error rate: 18.18% (min) Accepted variables pclass, title, & family.size determines/ predicts the survival of titanic passenger

4c. Applying Random forest and finding out the important variables for prediction

The titles (Mr, Mrs, Miss, Master) has value in predicting the passenger survival as compared to pclass (1, 2, 3), family size, and parch

4d.Visualizing as a decision tree

Titles of "Mr." and "Other" are predicted to perish at an overall accuracy rate of 83.2 %.

Titles of "Master.", "Miss.", and "Mrs." in 1st & 2nd class are predicted to survive

Titles of "Master.", "Miss.", and "Mrs." in 3rd class with family sizes equal to 5, 6, 8, & 11 are predicted to perish with 100% accuracy.

Titles of "Master.", "Miss.", and "Mrs." in 3rd class with family sizes not equal to 5, 6, 8, or 11 are predicted to survive with 59.6% accuracy.

If a person has title Mr and "other", he will survive at 91 instances.

If the title is neither Mr and "other", may be (Mrs, Miss), and if these people are in the 3 class, with alrge family size none of them will survive. Small family size, 83 instance of survival.

If the title is neither Mr and "other", may be (Mrs, Miss), and if these people are not in the 3 class, 168 instance of survival

5. Predicting variables for passenger survivals

variables *pclass, title, & family.size* determines/ predicts the survival of titanic passenger