IT Academy - Data Science with Python

S13 T01: Supervised Classification

Description

Let’s practice and become familiar with classification algorithms.

Level 1

• Exercise 1:

  • Create at least three different classification models to try to best predict DelayedFlights.csv flight delay (ArrDelay).

• Exercise 2:

  • Creates a new variable depending on whether the flight arrived late or not (ArrDelay> 0).

• Exercise 3:

  • Compare classification models using accuracy, a confidence matrix, and other more advanced metrics.

• Exercise 4:

  • Train them using the different parameters they support.

• Exercise 5:

  • Compare your performance using the traint / test approach or using all data (internal validation).

Level 2

• Exercise 5:
  • Perform some variable engineering process to improve prediction.

Level 3

• Exercise 6:
  • Do not use the DepDelay variable when making predictions.

Targets

• Classification trees
• KNN - k-Nearest Neighbors
• Logistic Regression
• Support Vector Machine
• XGboost