{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T17:44:56Z","timestamp":1782841496419,"version":"3.54.5"},"reference-count":0,"publisher":"ECMS","content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2026,6,23]]},"abstract":"<jats:p>Heterogeneity across electricity consumers poses a central challenge for short-term load forecasting (STLF), particularly when models must operate under realistic data and capacity constraints. This study investigates whether behavioral segmentation improves multi-meter forecasting accuracy compared with a single global model when total model capacity and feature space are strictly controlled. Using hourly Norwegian smart-meter data (342 meters across residential, industrial, and cabin categories), we evaluate six specialization strategies, including global learning, cluster-specific expert models, conditional shared representations, mixture-of-experts (MoE), hierarchical residual correction, and forecasting-optimized clustering. Results show that cluster-specific Random Forest experts achieve the best overall performance (MAE 0.8142 kWh), followed closely by a learnable MoE model, yielding modest improvements (1\u20132\\%) over a strong global baseline. However, per-meter normalization yields the largest gain (+4.5\\%), indicating that scale heterogeneity is the dominant modeling challenge. Improvements are most pronounced for industrial consumers, where structural diversity is highest. The findings suggest that careful heterogeneity management and moderate specialization are more impactful than increased model complexity in practical STLF deployment.<\/jats:p>","DOI":"10.7148\/2026-0568","type":"proceedings-article","created":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T16:55:14Z","timestamp":1782838514000},"page":"568-575","source":"Crossref","is-referenced-by-count":0,"title":["Cluster-based and expert specialization in short-term load forecasting: a capacity-matched empirical study"],"prefix":"10.7148","author":[{"given":"Saleh","family":"Alaliyat","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Rachid","family":"Oucheikh","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"4144","published-online":{"date-parts":[[2026,6,23]]},"event":{"name":"40th ECMS International Conference on Modelling and Simulation"},"container-title":["ECMS 2026 Proceedings edited by Filippo Sanfilippo, Florenc Demrozi, Fabio Sgarbossa, Mohammad Poursina"],"original-title":[],"deposited":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T16:55:20Z","timestamp":1782838520000},"score":1,"resource":{"primary":{"URL":"http:\/\/www.scs-europe.net\/dlib\/2026\/ecms2026acceptedpapers\/0568_simo_ecms2026_0075.pdf"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,6,23]]},"references-count":0,"URL":"https:\/\/doi.org\/10.7148\/2026-0568","relation":{},"subject":[],"published":{"date-parts":[[2026,6,23]]}}}