METODE ARTIFICIAL NEURAL NETWORK UNTUK MENENTUKAN AKURASI BOBOT PEKERJAAN PADA ESTIMASI BIAYA GEDUNG Studi Kasus : Gedung Dinas Pemerintah Kabupaten Trenggalek

Atho’illah, Muhammad Ibnu (2021) METODE ARTIFICIAL NEURAL NETWORK UNTUK MENENTUKAN AKURASI BOBOT PEKERJAAN PADA ESTIMASI BIAYA GEDUNG Studi Kasus : Gedung Dinas Pemerintah Kabupaten Trenggalek. Masters thesis, Untag 1945 Surabaya.

	Text Abstrak.pdf Download (381kB)
	Text Bab 1.pdf Download (136kB)
	Text Bab 2.pdf Download (233kB)
	Text Bab 3.pdf Download (560kB)
	Text Bab 4.pdf Restricted to Repository staff only Download (1MB)
	Text Bab 5.pdf Download (234kB)

Abstract

Requirements for State Buildings, in principle, must meet the administrative and technical requirements of buildings in general as regulated in UU 28 Year 2002 and its implementing regulations. However, as a building whose funding comes from APBN and / or APBD funds, it is necessary to fulfill additional / special requirements, such as funding documents, planning documents, development documents, and registration documents, in terms of administrative requirements, as well as building area standards. building classification, and certain technical specifications in terms of technical requirements. Cost estimation is generally required during the preparation stage of a project which is further broken down into the conceptual stage and the planning and consolidation stages. Cost is one of the main criteria in making decisions at an early stage during the building design design process. This study is devoted to providing an overview of the neural network method in assessing the work weight used in estimating the cost of government official buildings in the Trenggalek area. With the neural network method applied to previous project data, it is hoped that a special form of work weight from a budget plan can be obtained which can later be used to estimate similar projects in the Trenggalek area in general. The results of the analysis are obtained from comparing three modeling scenarios which are a series of parameters of building area, average column span, preparatory work, earthworks, plastering work, concrete work, floor and wall covering, door & window work, roof work,work painting, electrical work. From the three scenarios, the scenario with the best accuracy value in scenario 3 is selected with an MMRE accuracy value of 9.30%. This accuracy value is classified as good according to the AACE (Association for the Advancement of Computing in Education). The scenario has 9 variable architectural forms, 4 hidden layers and 1 target value which results in the smallest error in the 1st project of 0.44. %.

Actions (login required)

View Item