Effect of Proportion of Missing Data on Application of Data Imputation in Pavement Management Systems

Abstract

Missing data are commonly found in pavement condition/performance databases. A common practice today is to apply statistical imputation methods to replace the missing data with imputed values. It is thus important for pavement management decision makers to know the uncertainty and errors involved in the use of datasets with imputed values in their analysis. An equally important information of practical significance is the maximum allowable proportion of missing data (i.e. level of data missingness in the pavement condition/performance records) that will still produce results with acceptable magnitude of error or risk when using imputed data. This paper proposes a procedure for determining such useful information. A numerical example analyzing pavement roughness data is presented to demonstrate the procedure through evaluating the error and reliability characteristics of imputed data. The roughness data of three road sections were obtained from the LTPP database. From these data records, datasets with different proportions of missing data were randomly generated to study the effect of level of data missingness. The analysis shows that the errors of imputed data increased with the level of data missingness, and their magnitudes are significantly affected by the effect of pavement rehabilitation. On the application of data imputation in PMS, the study suggests that at 95% confidence level, 25% of missing data appears to be a reasonable allowable maximum limit for analyzing pavement roughness time series data not involving rehabilitation within the analysis period. When pavement rehabilitation occurs within the analysis period, the maximum proportion of imputed data should be limited to 15%.