Comparative evaluation of algorithms for GPS data imputation

Abstract

GPS data collection has been increasingly considered as an alternate means of data collection, replacing the traditional travel survey methods. Several algorithms which vary from informal ad-hoc approaches to advanced machine learning methods have been reported in the literature to extract activity and travel information from GPS traces. However, the differences in the performance of different algorithms are scarcely addressed. In this paper we evaluate the relative performance of different imputation algorithms for GPS data imputation by incorporating the naive Bayesian, Bayesian network, logistic regression, multilayer perception, support vector machine, decision table and C4.5. The accuracy of imputation results of various methods are compared using the GPS data collected in The Netherlands. Results show that the Bayesian network has a better performance than other algorithms according to the correctly identified instances and Kappa values for both training data and test data. Especially, the Bayesian network shows a stronger capability than other methods in the aspect of prediction generalisation.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 6 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.