Pregled bibliografske jedinice broj: 970723
Recursive Partitioning in Predicting Energy Consumption of Public Buildings
Recursive Partitioning in Predicting Energy Consumption of Public Buildings // Proceedings of the 29th Central European Conference on Information and Intelligent Systems / Strahonja, Vjeran ; Kirinić, Valentina (ur.).
Varaždin: Fakultet organizacije i informatike Sveučilišta u Zagrebu, 2018. str. 179-186 (predavanje, međunarodna recenzija, cjeloviti rad (in extenso), znanstveni)
CROSBI ID: 970723 Za ispravke kontaktirajte CROSBI podršku putem web obrasca
Naslov
Recursive Partitioning in Predicting Energy Consumption of Public Buildings
Autori
Zekić-Sušac, Marijana ; Has, Adela ; Mitrović, Saša
Vrsta, podvrsta i kategorija rada
Radovi u zbornicima skupova, cjeloviti rad (in extenso), znanstveni
Izvornik
Proceedings of the 29th Central European Conference on Information and Intelligent Systems
/ Strahonja, Vjeran ; Kirinić, Valentina - Varaždin : Fakultet organizacije i informatike Sveučilišta u Zagrebu, 2018, 179-186
Skup
29th Central European Conference on Information and Intelligent Systems (CECIIS 2018)
Mjesto i datum
Varaždin, Hrvatska, 19.09.2018. - 21.09.2018
Vrsta sudjelovanja
Predavanje
Vrsta recenzije
Međunarodna recenzija
Ključne riječi
Recursive partitioning, energy consumption, public buildings
Sažetak
Recursive partitioning includes a number of algorithms that create a classification or a regression decision tree by splitting the values of independent variables. The aim of this paper is to compare the accuracy of four different recursive partitioning methods in predicting the electrical energy consumption of public buildings. The input space included 141 attributes of public buildings in Croatia describing their geospatial, construction, heating, cooling, meteorological and energy characteristics. Four methods that produce regression tree partitioning were trained and tested. The results show that the random forest (RF) has outperformed CART, conditional inference tree (CTREE), and gradient boosted tree (GBT). The selection of important predictors was also compared and discussed.
Izvorni jezik
Engleski
Znanstvena područja
Ekonomija, Informacijske i komunikacijske znanosti
POVEZANOST RADA
Projekti:
HRZZ-IP-2016-06-8350 - Metodološki okvir za učinkovito upravljanje energijom s pomoću inteligentne podatkovne analitike (MERIDA) (Zekić-Sušac, Marijana, HRZZ - 2016-06) ( CroRIS)
Ustanove:
Ekonomski fakultet, Osijek