Files and links (1)
Published, Version of Record (VoR)CC BY V4.0, Open
Journal article
Machine learning approaches for predicting household transportation energy use
City and environment interactions, v 7, p100044
01 Aug 2020
Featured in Collection : UN Sustainable Development Goals @ Drexel
Abstract
This paper presents four modeling techniques for predicting household transportation energy consumption by exploring decision trees, random forest, and neural networks in addition to elastic net regularization analyses. The main objective of this study is to evaluate how effectively these advanced statistical models can be applicable to a Transportation Module (TM) operating within the Integrated Urban Metabolism Analysis Tool (IUMAT), a system-based computational platform for urban sustainability evaluation. The Delaware Valley Regional Planning Commission (DVRPC) travel demand model is used to estimate household transportation energy use based on household trip demand generation, travel mode, fuel type, distance and duration. The Household Travel Survey (HTS) and Traffic Analysis Zones (TAZ) drawn from the DVRPC database are used for model training. Our results indicate that machine learning algorithms, thanks to their ability to accommodate non-linearity, have significantly higher accuracy in predicting household transportation demand. We show that the Neural Network (NN) model out-performs the decision tree model, predicting transportation energy demand resulting in lower Mean Squared Error and a higher R-2. Using a Random Forest analysis for individual variable impact testing, we also demonstrate that the number of households' motorized trips and the travel distance are the most significant predictors of household transportation energy consumption.
Metrics
Details
- Title
- Machine learning approaches for predicting household transportation energy use
- Creators
- Shideh Shams Amiri - Drexel UniversityNariman Mostafavi - Drexel UniversityEarl Rusty Lee - University of DelawareSimi Hoque - Drexel University
- Publication Details
- City and environment interactions, v 7, p100044
- Publisher
- Elsevier
- Number of pages
- 9
- Grant note
- CBET 1554531 / National Science Foundation CAREER Grant; National Science Foundation (NSF)
- Resource Type
- Journal article
- Language
- English
- Academic Unit
- Civil, Architectural, and Environmental Engineering
- Web of Science ID
- WOS:000658365600001
- Scopus ID
- 2-s2.0-85104300592
- Other Identifier
- 991019167422704721
UN Sustainable Development Goals (SDGs)
This publication has contributed to the advancement of the following goals:
InCites Highlights
Data related to this publication, from InCites Benchmarking & Analytics tool:
- Collaboration types
- Domestic collaboration
- Web of Science research areas
- Environmental Sciences
- Environmental Studies
- Meteorology & Atmospheric Sciences