Neural Network for the Standard Unit Price of the Building Area. (2017)

• Record Type: Journal Article
• Authors: Fachrurrazi, ; Husin, Saiful; Tripoli, ; Mubarak,
• Is Part Of: Procedia engineering. Volume 171(2017)
• Page Start: 282
• Page End: 293
• Publication Date: 2017
• Subjects: Standard unit price -- HSBGN guidance -- building -- artificial neural network -- Indonesia -- Aceh; Engineering -- Congresses; Engineering -- Periodicals; Engineering; Conference proceedings; Periodicals; 620.005
• DOI: 10.1016/j.proeng.2017.01.336
• Languages: English
• ISSNs: 1877-7058

Abstract

Abstract: The standard unit price in the Unit Price of Public Building Construction ( Harga Satuan Bangunan Gedung Negara / HSBGN ), which is the guidance book for cost estimation of public building projects in Indonesia. This guidance has been implemented to estimate budget allocation for public building projects in Aceh. This guidance need to be reviewed regularly in order to adjust with the current condition. The review is carried out by comparing the standard unit price to the contractual unit price. The aim of this research is to develop a model for estimating standard unit price of building projects using Artificial Neural Network (ANN) and to review the standard based on the variables of inflation, interest rate and construction index that influence the cost of building projects. These variables are not completely independent but mutually interact one to another. ANN is the method that has ability to find the unique pattern through the learning process and to visualize the learning curve to solve the problem. Data from 156 contract of the public building projects have collected in this research. Review of the standard unit price of building projects in the guidance book shows 57.69% of the data is not distributed normally. The standard needs to be amended with an average correction 23.74%. The model for estimating standard unit price in this research has been generated based on MSE learning 0.000336 and MSE validation 0.008974.