چکیده مقاله
The residential sector is a primary source of greenhouse gas GHG emissions and plays a vital role in achieving emission reduction goals While most research focuses on annual CO۲ emissions for long term mitigation, predicting daily emissions is crucial for short term targets This study evaluates three machine learning models—Decision Tree DT , Random Forest RF , and Ridge Regression—in forecasting daily CO۲ emissions from residential buildings in the United States, a top polluting region Using data from ۱/۱/۲۰۲۲ to ۳۰/۱۲/۲۰۲۴ ۱,۰۹۵ days , the analysis split the dataset into ۸۷۵ training days January ۲۰۲۲ to May ۲۰۲۴ and ۲۱۹ testing days May ۲۰۲۴ to December ۲۰۲۴ Ridge Regression outperformed the others, achieving the lowest error RMSE: ۰ ۰۹۴۲ and highest accuracy R²: ۰ ۹۵۹۶ , capturing ۹۶% of emissions variation Random Forest followed closely RMSE: ۰ ۰۹۶۷, R²: ۰ ۹۵۷۵ , while Decision Tree performed significantly worse RMSE: ۰ ۱۲۵۸, R²: ۰ ۹۲۸۰ The high R² values >۰ ۹۲ across all models highlight the effectiveness of temporal features in predicting emissions patterns Ridge Regression’s superiority suggests residential CO۲ emissions follow a linear trend, with regularization preventing overfitting, aligning with seasonal and yearly trends These findings demonstrate the potential of machine learning, particularly Ridge Regression, in accurately forecasting daily CO۲ emissions from residential buildings, aiding policymakers in setting and monitoring short term emission reduction targets
کلیدواژهها
نویسندگان
Maryam Pournaghi Keykele
Master of construction management Science and research university, Tehran, Iran
مهدی روانشادنیا
Associate professor, Science and research university, Tehran, Iran
شیوه ارجاع
Pournaghi Keykele، Maryam و Ravanshadnia، Mehdi،1403،Comparative Analysis of Three Machine Learning Models for Predicting Daily CO۲ Emissions in the Residential Sector: A Future Trend Forecasting Approach،هشتمین کنفرانس بین المللی عمران، معماری، شهرسازی با رویکرد توسعه زیرساخت های شهر�
ارائهشده در
مجموعه مقالات نهمین کنفرانس بین المللی عمران، معماری، شهرسازی با رویکرد توسعه زیرساخت های شهری30 مرداد 1404