Defining Contractor Performance Levels

In many countries, state and private owners often categorize contractors into various groups based on their capacities as well as their previous performance. These categories are then used for several purposes including contract award and prequalification. A number of different statistical algorithms can be used to categorize the contractors. No one algorithm can be used globally in all situations since the performance of these algorithms depend partially on the actual data set being used. Therefore this paper presents a new model for grouping contractors based on historic data. The model developed here utilizes qualitative as well as quantitative measures about the contractors’ performance. The model addresses the issue of classifying contractors into various performance categories using different crisp and fuzzy clustering algorithms and assesses the performance of these algorithms with appropriate validity measures. The model was validated using actual data from previously recorded project information for 13 contractors. The analysis shows that the model can be used effectively to determine the performance level of the contractors and to perform clustering of contractors into different performance groups.