Skip to content

Latest commit

 

History

History
27 lines (19 loc) · 1.12 KB

README.md

File metadata and controls

27 lines (19 loc) · 1.12 KB
Raw

Titanic Passenger Survival Prediction

SKLearn KneighborsClassifier model that predicts a Titanic passenger's survival

Data

My class provided a dataset with information about the titanic passengers and their info. Features in the dataset: Survival, Ticket class, Sex, Age in years, # of family members, Ticket number, Passenger fare, Cabin number, and Port of Embarkation.

Data Processing

I dropped all n/a values as well as multiple features that didn't correlate to a passenger's survival. As well as some data mapping and rounding floats to ints.

EDA

EDA was performed by creating graphs with matplotlib and seaborn. Graphs include visualing the correlation of data and class balances.

Model

Train and test data was created using SKLearn's train_test_split() Model is trained and fitted using SKLearn's KneighborsClassifier.

Validation

Model is scored on test and train using SKLearn's score() A cross validation score is obtained using SKLearn's cross_val_score() A classification report is obtained using SKLearn's classification_report() A heatmap of the confusion matrix from the model's prediction is created using seaborn.