000			03060nam a2200301 i 4500
003			CSPC
005			20261009091603.0
008			261009s2026 ph a b 000 0 eng d
040			_aCSPC _beng _cCSPC _erda
090			_aMTH ME G984m 2026
100	1		_aGutierrez, Rosalie I., _eauthor.
245	1	0	_aMachine learning-driven prediction of electrical consumption in higher education institutions: a comparative analysis incorporating occupancy dynamics at Camarines Sur Polytechnic Colleges / _cRosalie I. Gutierrez.
300			_a133 leaves.
336			_2rdacontent _atext
337			_2rdamedia _aunmediated
338			_2rdacarrier _avolume
502			_aMaster's Thesis (ME) - Engineering. _cCamarines Sur Polytechnic Colleges, 2026.
504			_aIncludes bibliographic references.
520	3		_aHigher education institutions (HEIs) consume substantial energy, often comparable to small cities, due to operations such as lighting, HVAC systems, and electronic equipment. In the Philippines, energy costs are a major burden for public institutions, making efficiency essential for fiscal sustainability and compliance with the Energy Efficiency and Conservation Act of 2019. Camarines Sur Polytechnic Colleges (CSPC), a state-run institution in the Bicol Region, exemplifies these challenges, with consumption patterns shaped by student and employee populations, seasonal variations, and external shocks such as the COVID-19 pandemic. This study applies supervised machine learning models—Linear Regression, Random Forest, and Gradient Boosting—to predict CSPC’s monthly electricity consumption. Input features include student and employee population data, along with engineered variables such as pandemic indicators and seasonal flags, addressing gaps in traditional forecasting methods that often overlook occupancy dynamics. Analysis identified normalized year (0.2176), employee numbers (0.2036), and total occupancy (0.1632) as the primary drivers of CSPC’s energy use, accounting for 58.44% of the Random Forest model’s predictive power. Random Forest emerged as the most reliable model, achieving a testing of 0.6995. explaining nearly 70% of consumption variability with stable performance across data samples. Forecasts projected 289,495 kWh for late 2024 and 655,068 kWh for 2025, with peaks during enrollment months and lows in vacation periods. By integrating population dynamics and external factors, this study demonstrates the potential of machine learning for proactive energy management in HEIs. The findings support global efforts toward energy efficiency and align with the United Nations Sustainable Development Goal 7: Affordable and Clean Energy.
650			_aMachine learning.
650			_aHigher education institutions.
720			_aCerillo, Jose Eduardo B., _eadviser.
720			_aPontillas, Virginia V., _echair, panelist.
720			_aPacer, Wenifredo L., _epanelist.
720			_aDematera, Fernan D., _epanelist.
942			_2ddc _n0 _cUM _hG984m _iME _kMTH _m2026
999			_c33665 _d33665