ConForMiSt: A Multi-model Dual-phase Framework utilizing Machine Learning for Carbon Footprint Prediction and Reinforcement Learning for decision optimization

Abstract

Over the past decade, there has been a significant surge in harmful waste emissions of greenhouse gases namely carbon dioxide, methane and fluorinated gases in the atmosphere. Two major categories of activities can be broadly identified which have contributed to this condition. The first is proliferation of world- wide industrial activity accounted by the industrial plants across all the major continents. Second is the human activity which also contributes to carbon emissions produced as a result of wide-ranging everyday activities that involve use of electricity, transportation, food consumption and other consumer-mindset driven activities. This article focuses on the second category to assess, evaluate and recommend suitable mitigation measures to regulate usage patterns. A system is conceptualized and built to gather carbon emission data. The data gathered by the system is analyzed to detect patterns in carbon footprint generation. The behavior of carbon footprint patterns is subsequently analyzed using mathematical models. Post-analysis, machine learning models are leveraged to make predictions of carbon footprint. A reinforcement learning framework is proposed to analyze predictions and generate recommendations, taking into account user preferences. A web application is developed to visualize various aspects like usage patterns, predictions, and recommendations on a personalized level. These aspects are then utilized to provide insights at the aggregated level in the context of a group of individuals. The scope of the proposed methodology can be extended to a broader group of stakeholders to ensure reduction in carbon footprint emissions at higher aggregated levels.