Developing an optimised activity type annotation method based on classification accuracy and entropy indices

摘要: ABSTRACTThe generation of substantial amounts travel- and mobility-related data has spawned the emergence era big data. However, this generally lacks activity-travel information such as trip purpose. This deficiency led to development purpose inference (activity type imputation/annotation) techniques, which performance depends on available input (number of) activity classes infer. Aggregating types strongly increases accuracy is usually left discretion researcher. As open for interpretation, it undermines reported accuracy. study developed an optimised classification methodology by identifying with optimal balance between improving model accuracy, preserving from original set. A sensitivity analysis was performed. Additionally, several machine learning algorithms are experimented with. The proposed method may be app...