Abstract:
Cost estimating in construction is critical at the early stage of the project in order to
determine whether the project is feasible or not to the client. Cost estimators are facing
numerous difficult moments due to incomplete project details and unavailability of cost
estimation instruments in early stages of a project. This issue uplifts the necessity of
novel and advanced cost model which would be simple, understandable and more
reliable. Within the Sri Lankan context, the accuracy of the estimated amount is solely
based on the experience and skills of the estimator due to unavailability of reliable cost
estimation tool for educational institution buildings in Sri Lanka. Therefore, this paper
aims to discuss about the development of a computer model for cost estimation in
educational buildings with the current practices and related issues in preparing
preliminary project estimates. Semi structured interviews were conducted between
twelve experts from consultancy and contractor organisations and the necessary cost
data were selected from twenty educational buildings. The Multiple Regression Analysis
and Artificial Neural Network methods were utilised to analyse the collected data. Each
method has a unique way of building relationships between predictors and responses.
However, both the methods were succeeded only in estimating cost of limited number of
sub-elements. Multiple Regression Analysis succeeded on five occasions and Artificial
Neural Network method had presented efficacy in seven sub-elements only. Altogether
eight elements were succeeded in estimating the cost of an educational institution
buildings.
