Introduction to data and features

This week we will look a bit more closely at an all-important component of any machine learning model, the input data and associated features.

In particular we will cover:

• types of data and features, and feature selection
• feature extraction from raw data – specifically from images
• labelling image data – what is in your images and where
• pre-processing of data – how to manage imperfect datasets.

At the end of the week there will be a practical in which you will use what you’ve learned to open a set of image data, perform some pre-processing, and extract some image data.

We’ll give you the full instructions in the practical, and the data we will use is a subset of the publicly available dataset of images of flowers provided by the Visual Geometry Group at the University of Oxford.

For more information on the dataset see Flower data.

We will use a subset of the flower dataset hosted on our Github repository PhenoDataCampp.