Soil Classification using Machine Learning: From First Principles
Published:
Brief Overview
The term project (AI for Engineers course @ VNIT Nagpur), tries to harnessing the power of Artificial Neural Networks (ANNs) and Landsat satellite data to classify soils. The main focus being understanding the “Black-Box” nature of ANNs.
Why “Unveiling the Black Box”?
This approach isn’t just about achieving accurate soil classification; it’s about demystifying the black box nature of Artificial Neural Networks. The simplicity of this dataset enables to compare the ANN output to the well studied statistical approaches like Nearest Neighbour or KNN Classifiers.
Results
The results were synthesized into an interactive web app which can be found below.