search

odysszis / AML

Repository for kaggle projects for the module Applied Machine Learning
0 stars 0 forks source link

Complete 'LoadData' and 'PreProcess' data methods for Data Super Bowl #2

Closed PeadarOhAodha closed 8 years ago

PeadarOhAodha commented 8 years ago

The very basic methods required to get the model prepped for modelling steps.

There will be lots of scope for sophisticating on any pre processing steps at a later point, so implement as a class with with the ability to extend associated methods later.

MaxHoefl commented 8 years ago

Note the data pre-processing chapter in "modeloutline.pdf" in the 3rd issue on waffle

PeadarOhAodha commented 8 years ago

Sunny brook data: Serialised numpy arrays for both the resized training and corresponding mask arrays [numimages* height * width](first image in array corresponds to first mask, etc) pushed to git. Use numpy.load() to access and matplotlib to visualise. Awaiting full kaggle provided un masked training set to run LoadData.py.