Project: Predictive Modeling with Azure Machine Learning Studio

Taught by: Snehan Kekre, Machine Learning Instructor, Machine Learning

Introduction

In this project, I learned to use Azure Machine Learning Studio to build a predictive model without writing a single line of code. I predict flight delays using weather data provided by the US Bureau of Transportation Statistics and the National Oceanic and Atmospheric Association (NOAA).

Main Objectives

• Build a predictive model using Azure ML Studio.
• Demonstrate a working knowledge of setting up experiments on Azure ML Studio.
• Operationalise machine learning work flows with Azure's drag-and-drop modules.

Step Process

• Importing the Data Sets
• Scrubbing Missing Values
• Eliminating Target Leaks
• Conversion to Categorical Features
• Preparing Features to be Joined with Weather Data
• Preprocessing the Weather Dataset
• Joining Both Datasets
• Training and Evaluating the Model

Result:

Accuracy: 76.9%

• Figure: ROC Curve

• Summary of the model

Accomplishment

• Applied Two-class logistic Regression to predict the model
• Able to train and evaluate a predictive model on Azure Machine Learning Studio, all without writing a single line of code!
• Able to predict flight delays using weather data provided by the US Bureau of Transportation Statistics and the National Oceanic and Atmospheric Association (NOAA).