Fang, Z; Pitt, M; Hanna, S; (2019) Machine learning in facilities & asset management. In: Proceedings of the 25th annual Pacific rim real estate society (PRRES) conference. PRRES: Melbourne, Australia.

Preview

Text Hanna PRRES2019_Full_Fang_Amended_For_Submision.pdf - Accepted Version Download (804kB) | Preview

Abstract

In this article, we explore a machine learning approach that helps the Facility Management (FM) manager and FM data analyst to do the FM data clustering and classification automatically. Through experiment the popular machine learning algorithm, we examine how the current available Machine Learning (ML) & Natural Language Processing (NLP) technic can help solve the interoperability issue faced in the field of FM. The finding of this research indicates that: 1. The deep learning network is able to classify the building assets according to their group elements (level three of NRM’s Elemental Standard of Cost Analysis) with a high accuracy rate (by more than ninety percent accuracy);2. The Convolutional Neural Network (CNN) Classifier can achieve better accuracy performance than the junior building data analyst; 3. The Unsupervised Skip-gram Gradient Descent Model can cluster the words in the document into different groups; 4. The Unsupervised Skip-gram Gradient Descent Model can reveal the hidden relationship inside the FM data. For future researches and projects, this research enlightens the future direction of applying ML/NLP techniques in the field of FM.

Download activity - last month

Export as CSVXMLJSON

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as JSONXMLCSV

Archive Staff Only

View Item