Vehicle Dimensions Based Passenger Car Classification using Fuzzy and Non-Fuzzy Clustering Methods

There has been globally continuous growth in passenger car sizes and types over the past few decades. To assess development of vehicular specifications this context to evaluate changes powertrain technologies depending on surrounding frame conditions, such as charging stations vehicle taxation policy, we need a detailed understanding fleet composition. This paper aims therefore introduce novel mathematical approach segment vehicles based dimensions features using means fuzzy clustering algorithm, Fuzzy C-means (FCM), non-fuzzy K-means (KM). We analyze performance proposed algorithms compare them with Swiss expert segmentation. Experiments real data sets demonstrate that FCM classifier better correlation segmentation than KM. Furthermore, outputs from five clusters show algorithm superior for accurate categorization because its capacity recognize consolidate dimension attributes unsupervised set. Its categorizing was promising an average accuracy rate 79% positive predictive value 75%.