A two-staged approach for assessing for the quality of internet survey data

In this work, we propose to develop a procedure to detect eirors in data collected through Internet surveys. Although several approaches have been developed, they suffer many limitations. For instance, many approaches require prior knowledge of data and hence they need different procedures for different applications. Others have to test a large number of parameter values and hence they are not very efficient. To develop a procedure to overcome the limitations of existing approaches, we by to understand the nature of this quality problem and establish its linkage to travelling salesman problem (TSP). Based on the TSP problem structure, we propose to develop a two-staged approach based on a genetic algorithm to help ensure the quality of Internet survey data.