Human daily activity behavioural clustering from Time Use Survey

Abstract

Identification of daily pattern behaviours of people from Time use Survey with the purpose of defining archetypes of persons is becoming a new rising research field. Identified patters are useful for developing more realistic models to simulate activities of citizens related to mobility and households energy consumption. These models are required to test and develop simulation scenarios of future smart grids and cities. In this work we apply the k-modes algorithm to clusterize the Italian TUS data-set. For the best of our knowledge this is the only study that applied unsupervised clusterization and classification of Italian TUS data and the only one that extended the analysis to mobility activities of the TUS data-sets. From experimental results we obtained different clusters for weekdays, saturdays and holidays, respectively.