How to Optimize Gower Distance Weights for the k-Medoids Clustering Algorithm to Obtain Mobility Profiles of the Swiss Population

Abstract

This piece of research aims to obtain mobility profiles of the Swiss population. To that end, a survey of the Swiss Statistical Office (FSO) called Mobility and Transport Micro-census (MTMC) is utilized. Along with a qualitative method clustering, the respondents in the survey are clustered based on their mobility characteristics to obtain their profiles. The clustering, in particular acquiring medoids (centrotypes or exemplars), helps us then to generate a synthetic population of Switzerland. To gain medoids of each cluster, the k-Medoids clustering algorithm is utilized which partitions instances based on their positions in a latent space (symmetric distance matrix). Distances that shape this space can be generated by various metrics e.g. Euclidean, Gower, Manhattan. Since in this study features are mixed-type (e.g. numeric, categorical, etc.), the Gower distance metric is preferred. In this study, the default weights of the Gower distance are optimized to obtain a higher Average Silhouette Width (ASW) value of the clustering results. ASW can be used to measure the quality of clustering results in which high value leads to higher intra-cluster homogeneity and inter-cluster dissimilarity. So, maximizing the ASW value improves the quality of the clusters which is the goal of the optimization. At the end, this process helps us to obtain more accurate mobility profiles of the Swiss population.

A meso-level empirical validation approach for agent-based computational economic models drawing on micro-data: a use case with a mobility mode-choice model

“…As for the clustering algorithm, we utilized the k-medoids clustering algorithm, since it is compatible with the latent space created by the Gower distance metric (Bektas and Schumann 2019a). However, k-medoids is an unsupervised algorithm; thus, we need to find ex-ante the optimal number of clusters.…”

“…There are the goodness-of-fit metrics in the literature [e.g., Average Silhouette Width (ASW), Calinski and Harabasz Index (CH) and Pearson version of Hubert's Γ (PH) (Campello and Hruschka 2006)], which can provide quantitative measurement scores regarding the quality of clustering with the different number of clusters. The ASW is one of the most widely used approaches that measures how well an instance is matched with its own cluster (Maulik and Bandyopadhyay 2002;Bektas and Schumann 2019a). As a goodness-of-fit measure, it reflects how well intra-cluster homogeneity and intercluster dissimilarity are maximized (Rousseeuw 1987).…”

“…As ex-post behavior, mode-choice should be in the chosen variables (the last variable in the list below). As for the variables constituting ex-ante characteristics, we identified the ones influencing mode-choice in our previous research (Bektas and Schumann 2019a) and use them in this case study (the first four variables in the list below). The full list of the used variables is demonstrated below:…”

“…The supporting information regarding the data source can be found in Bektas and Schumann (2019a) and ARE/BfS (2017).…”

A Socio-psychological Approach to Simulate Trust and Reputation in Modal Choices

“…In the configuration phase, we applied cluster analysis on the Swiss Household Energy Demand Survey (SHEDS) [36] and the Mobility and Transport Micro-census (MTMC) [2] to generate the mobility profiles. This process is detailed in [4]. A number of random respondents from the surveys that matched these clusters are selected to generate a synthetic agent population for simulation.…”