Mastery Level

Predict the survival of Titanic passengers from the file $\text{'}$~$/$nclab$-$data$-$read$/$titanic$-$test.csv$\text{'}$ based $($solely$)$ on their fare!

The program below reads the training data file and fits the model. It's up to you to make the prediction. # Import Pandas and Statsmodels:

import pandas as pd$,$ statsmodels.api as sm

# Read CSV data file:

df $=$ pd$.$read$\_$csv$(\text{'}$~$/$nclab$-$data$-$read$/$titanic$.$csv$\text{'})$

# Define the training data X and Y:

X $=$ df$[\text{'}$Fare$\text{'}]$

Y $=$ df$[\text{'}$Survived$\text{'}]$

# Add constant to X and store the result in a new array Xc:

Xc $=$ sm$.$add$\_$constant$($X$)$

# Fit the logistic model with Statsmodels:

model $=$ sm$.$Logit$($Y$,$ Xc$).$fit$()$

# Read the testing data file into a new DataFrame df$\_$test:

df$\_$test $=$ pd$.$read$\_$csv$(\text{'}$~$/$nclab$-$data$-$read$/$titanic$-$test.csv$\text{'})$

# Define the testing dataset F $($column 'Fare' from df$\_$test$)$:

F $=$ df$\_$test$[\text{'}$Fare$\text{'}]$

# Add constant to F and store the result in a new array Fc:

Fc $=$ sm$.$add$\_$constant$($F$)$

# Make the prediction using Fc and store the result in Prob:

Prob $=$ model.predict$($Fc$)$

# Create a list 'Surv' of PassengerIds of all passengers whose

# probability of survival is at least $0\mathrm{.}5$:

Surv $=$ list$($df$\_$test$[\text{'}$PassengerId$\text{'}].$loc$[$Prob$]>=0\mathrm{.}5)$

# Sort the list Surv:

Surv.sort$()$

# Main program $($do not change$)$:

print$("$PassengerIds of predicted survivors:

$"+$ str$($Surv$))$

Use the comments in the code as a guide. The main program will then print the PassengerIds of the predicted survivors.

Note: As we already stated in $5\mathrm{.}7,$ considering the fare as the only explanatory variable is a strong simplification.