Comparison of the Performance of K-Nearest Neighbours and Generalized Neural Network in Construction Crew Productivity Prediction

Construction crew productivity prediction is one of the most important issues that affect realistic construction duration and cost. Use different search algorithms like Feed Forward Neural Network, Ant Colony, Artificial Bee Particle Swarm Optimization, Radial Based Networks Self Organizing Maps for problem have been discussed in previous studies. However, significant effect coherence between nature data characteristics method used performance has generally neglected. The aim current research thus to analyse two contemporary learning algorithms; K- Nearest Neighbour (K-NN) Generalized Network (GRNN) when applied three (formwork, tiling masonry) related sets with distribution characteristics. Performance both methods varied changing coefficient variation values. K-NN outperformed GRNN all had their worst on dataset highest variance.