Bidding is a commonly used method by
contractors to secure business for their companies. A typical bid price
consists of three main cost components; direct costs, indirect costs and
mark-up. While there are methods for estimating direct and indirect
costs with or without contingency, there are fewer tools for estimating
mark-ups. This study introduces Multiple Regression (MR), Artificial
Neural Network (ANN) and Adaptive Nero-Fuzzy Inference System (ANFIS)
techniques for estimating markups, utilizing data collected from
contractors in Canada and the USA. To facilitate the modeling process,
thirty factors, recognized to impact the mark-ups of estimates and 23 of
which are clustered in into five independent categories: need for work,
job uncertainty, job complexity, market condition, and owner
capability. The ANN, ANFIS and MR models are developed using “MATLAB
2017a” and the results obtained by these methods are compared to their
respective performance against the actual estimated markups by the
contractors. The ANFIS model yielded more accurate estimates than the
other two (2) models.