Weighted score-driven fuzzy clustering of time series with a financial application

Multiway clustering with time-varying parameters

“…However, as shown in Cerqueti et al. ( 2022 ), it could be interesting to assign different weights to the distribution parameters and search for the optimal weights. This aspect should be taken into account in future studies.…”

“…Indeed, in the present paper we implicitly assign equal weights to the different time-varying parameters. However, as shown in Cerqueti et al (2022), it could be interesting to assign different weights to the distribution parameters and search for the optimal weights. This aspect should be taken into account in future studies.…”

“…Despite clustering techniques, based on time series’ distribution characteristics, have been extensively studied, approaches based on time-varying parameters have only recently been explored in Cerqueti et al. ( 2021 , 2022 ).…”

“…The idea of considering when alternative features have their relevance in grouping the time series. Cerqueti et al (2022) overcome the problem related to the selection of the target parameter by using more parameters jointly, focusing on the use of unconditional and conditional quantities in the clustering process. We have to acknowledge that the proposed unconditional distribution-based clustering provides results that are very close to the static parameters' ones, even if the clustering interpretation is much more interesting.…”

Accelerated univariate sequential clustering

“…A clustering algorithm aims to group objects into similar clusters to minimize the total cost (i.e., the measure of dissimilarity). For data clustering, a number of methods have been developed, including hierarchical clustering [14], K-means, self-organizing maps [15], fuzzy clustering [16], and nearest neighbor clustering algorithms [17].…”