Financial Services Advisory

Scope Emissions Prediction Model Linear Models vs Machine Learning Approach

By:
insight featured image
How Grant Thornton can help you to estimate scope emissions for companies that do not report these metrics and incorporate these into banks’ Climate & Risk Quantification framework.
Contents
Scope Emissions Prediction Model Linear Models vs Machine Learning Approach

Scope Emissions Prediction Model Linear Models vs Machine Learning Approach

Read our full publication.
Download PDF [5529 kb]

In this paper, we focus on the development of a model that predicts scope emissions for any company that cannot currently provide this information. Our model is built on external data and estimates relationships between these corporate entities’ financial information, sector, region and other information, and scope emissions reported by these entities.

The output of the model can be used by both individual entities as well as banks and other financial institutions in order to estimate the emissions structure of their portfolios. In our methodology we compare simple linear models as well as more advanced machine learning techniques.

The best model estimates are achieved using a Linear Mixed Models Methodology.

 

Asset 1.png

Asset 2.png

 

 

The publication is structured as follows: In the first section, we outline key modelling methodologies considered for the model build. After this, we outline the model build process, key modelling inputs and outputs. The third section aims to select the best model for emission prediction. In the final section, we focus on the practical implementation and deployment of this model. We also seek to conclude and point out potential issues and areas for further development.