Data-Driven Solutions for Single-Stream Recycling Optimization (A Case Study)

My capstone project is a case study on Data-Driven Solutions for Single-Stream Recycling Optimization. My aim is to describe the numerous ways that Big Data and Data Science are being used to aid solutions for reducing the amount of municipal solid waste (MSW) and optimizing recycling.  There are many interesting Artificial Intelligence and Machine Learning concepts that are being used by different organizations to combat this problem.  I plan to go into depth on how modern computer vision techniques like convolutional neural networks (deep learning) can be used for automated recycling classification. My primary objective is to identify a “best” machine learning modeling technique to use for recycling segregation in Materials Recovery Facilities (MRFs).  I use an open-source dataset called TrashNet to tune, train and test the candidate models. My secondary objective is to assess the model’s ability to predict the classification of an item quickly and decide if it is viable for use at an MRF.