FUNDAMENTALS OF DEVELOPING CONCEPTUAL COST ESTIMATION MODELS USING MACHINE LEARNING TECHNIQUES - SELECTION AND MEASUREMENT OF BUILDING ATTRIBUTES

Aug 4, 2026

Upload by Oliver

Ensuring the identification of building attributes is central to developing machine learning cost estimation models, yet prior work remains limited in two ways: it does not consistently categorize building characteristics by cost type, and it fails to define clear measurement standards for attribute qualities. This study selects a building-attribute set specifically for conceptual cost estimation and for establishing measurement standards. A two-round focused-group discussion process yields 13 attributes collectible before design completion, supporting model development assessment and application-phase evaluation.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages15

Reading Time38 min

AccessFree to download

Abstract

Introduction

Selection and Measurement of Building Attributes

Literature Review

PLANNING MALAYSIA: Journal of the Malaysian Institute of Planners VOLUME 22 ISSUE 3 (2024), Page 242 – 256 FUNDAMENTALS OF DEVELOPING CONCEPTUAL COST ESTIMATION MODELS USING MACHINE LEARNING TECHNIQUES: SELECTION AND MEASUREMENT OF BUILDING ATTRIBUTES Rui Wang1, Hafez Salleh2, Zulkiflee Abdul-Samad3, Nabilah Filzah Mohd Radzuan4 and Kok Ching Wen5 1,2,3 Centre of Building Construction and Tropical Architecture (BuCTA), Faculty of Built Environment, UNIVERSITI MALAYA 4 Faculty of Computing, UNIVERSITI MALAYSIA PAHANG 5 Faculty of Engineering and Quantifying Surveying, UNIVERSITI ANTARABANGSA INTI Abstract Ensuring the identification of building attributes is the primary task in developing a machine learning cost estimation model. However, the existing research on building attributes has the following shortcomings: it struggles to categorize building characteristics according to various cost types, and the suggested sets of attributes do not clearly establish measurement standards for these qualities. To address these issues, this study aims to select a set of building attributes suitable for conceptual cost estimation and establishment of measurement standards. Through a two-round process of focused group discussions, this research ultimately identified 13 building attributes that can be collected before the completion of building design. These attributes serve as a basis for assessing completed building projects during the model development phase and for evaluating new projects during the model application phase. This study provides a foundational framework for the development of conceptual cost estimation models, ultimately enhancing the accuracy of machine learning cost estimation models. Keywords: Conceptual cost estimation, machine learning, building attributes INTRODUCTION The success of construction projects is heavily dependent on cost prediction (Juszczyk, 2020; Park et al., 2022; B. Wang et al., 2021) . A successful construction project should achieve on-time delivery within the budget while yielding a substantial return on investment (Ma et al., 2016; Peleskei et al., 2015) . Inaccurate estimations can lead to cost overruns, construction delays, and many other, even worse, outcomes (Car-Puši & Mladen, 2020; Elmousalami, 2020; Mir et al., 2021) . Currently, qualitative and quantitative analyses are the primary methods for cost estimation (Hashemi et al., 2020). Qualitative methods based on expert judgment may introduce biases and result in inaccurate estimates (R. Wanget al., 2022) . Quantitative techniques not only rely on historical data and expert knowledge but can also analyze project design, processes, and unique characteristics (Ugur, 2017) . Quantity surveying is considered the most reliable quantitative method for obtaining construction costs (Ugur et al., 2018) . However, it requires surveyors to possess substantial expertise, can be timeconsuming, and is only feasible with a well-developed design (Ugur et al., 2018) . Therefore, there is a need to enhance the efficiency of quantity surveying. Other traditional quantitative cost estimation methods mostly depend on statistical analysis and simple regression theory, which results in lower accuracy and longer time consumption and does not add value to cost estimation (Jiang, 2019) . As a result, traditional budgeting methods no longer meet the needs of practical engineering budgets. It is essential to use computer technology for intelligent cost control in construction budgeting to improve accuracy (Abdel-Basset et al., 2020; Patil & Salunkhe, 2020; Xuan & Li, 2022) . Machine Learning (ML) is a branch of Artificial Intelligence (AI) and is a data-driven modelling technique (Brink et al., 2016) . It can be used to automatically extract hidden patterns from high-dimensional data and convert them into explicit information or knowledge to address challenging issues in the construction industry (Zhou et al., 2018) . As the name suggest

FAQs

What is the primary focus of developing a machine learning cost estimation model?

Identifying building attributes. The study treats standardized attribute selection as the key prerequisite for building effective conceptual cost estimation models.

What shortcomings in existing building-attribute research does the study address?

The research is limited in categorizing building characteristics by different cost types and in providing clear measurement standards for both numeric and textual attributes.

How were the building attributes selected and what result was produced?

A two-round focused group discussion builds on a prior standardized attribute set and results in 13 attributes that can be collected before design completion.

Aug 4, 2026

Upload by Oliver