use again the ge flight quest data and build on


Use again the GE Flight Quest Data and build on the work done for homework 1 and 2. The goal is to create a classification model which classifies flights into "on-time" (you can define how many minuteslate you still accept) and "delayed". Use all available data (there is weather information available). Notealso that you have data for several days which you can use for training and testing.

Then, add 4 columns for the delay. These delay are:

  • gate_departure_delays =>(actual_gate_departure-scheduled_gate_departure )
  • gate_arrival_delays =>(actual_gate_arrival-scheduled_gate_arrival)
  • runway_departure_delays =>(actual_runway_departure-scheduled_runway_departure)
  • runway_arrival_delays =>(actual_runway_arrival-scheduled_runway_arrival)

Then work on this file.

Follow the CRISP-DM framework

3. Data Preparation

  • Construct variables as needed.
  • Deal with missing data.

Depend on the model that you use, how it deals with messing data, such as replace them with the mean values or find out what values should be, or anything you comfortable with, and provide a short description for that.

  • Combine the flight history data with additional data (weather, etc.)

Add column for weather to the file history that we have. You can find weather data in the data files.

4. Modeling

  • Create at least 3 different classification models (different techniques) and discuss theadvantages of each model for this classification task.

They could be 3 different trees, and describe each model. For example, put the tree or part of it then explain everything "what does it mean? What it tell the airport? What does it mean for them? And what can they do?

  • Assess how well each model performs (use training/test data, cross validation, etc. asappropriate).

Which of these models work better?

5. Evaluation

  • How useful is your best model for the airline industry? How would you measure the model'svalue?

6. Deployment

  • How would you deploy the model?
  • Write a report covering in detail all steps of the project. The results have to be reproducibleusing this documentation.

Request for Solution File

Ask an Expert for Answer!!
Database Management System: use again the ge flight quest data and build on
Reference No:- TGS0264768

Expected delivery within 24 Hours