B1799
Title: Using a two-step clustering approach to examine courts' efficiency in European countries
Authors: Maria Stachova - Faculty of Economics, Matej Bel University in Banska Bystrica (Slovakia) [presenting]
Jan Hunady - Matej Bel University (Slovakia)
Abstract: Panel data, or longitudinal data are collected and analyzed in different fields of research areas. This type of data contains statistical objects that are periodically observed over time. Compared to cross-sectional data, the number of clustering techniques suitable for panel data is significantly limited. It is why, the main goal of the contribution is to present a two-step clustering approach, where in the first step, the panel data are transformed into a static form using a set of proposed characteristics of time dynamic. In the second step, the objects are clustered by conventional spatial clustering algorithms, such as K-means clustering or hierarchical partitioning. The clustering performance of the mentioned approach is then compared with two extant methods using real panel data sets. Data consists of indicators capturing the effectiveness of courts at the level of the first instance. The used methodology allowed us to group European countries based on the efficiency of their courts as well as to capture the dynamic trends. This approach can be in general helpful for assessing and comparing the efficiency of public finance spending and assessing the quality of public institutions including courts.
