Last modified: 2024-05-14
Abstract
A person $X$'s CV, $ CV_X = ( e_1, e_2, \ldots, e_{k_x}) $, consists of a sequence of events $e_i$. For an event $ e_i = (s_i, f_i, R_i, S_i, \ldots) $ we at least know its start date $s_i$, its end (finish) date $f_i$, the type $R_i$ of the event, the state (location) $S_i$ of the event, and maybe something more. We decided to base our analysis on the corresponding \emph{co-presence network} -- a weighted multi-relational temporal network $N = (V, L, w, t)$ in which the set of \emph{nodes} $V$ consists of studied persons. There is a \emph{link} (edge) $\ell = (u:v; R)$ of type (relation) $R$ between persons $u, v \in V$ iff there exist events $e_u = (s_u, f_u, R_u, S_u)$ and $e_v = (s_v, f_v, R_v, S_v)$ such that $R_u = R_v = R$ and $S_u = S_v$ and $[s, f] = [s_u, f_u] \cap [s_v, f_v] \ne \emptyset$. The weight of the link $\ell \in L$ is $w(\ell) = f - s$, that is the length of the corresponding time interval $[s, f]$. The traditional sequence analysis deals mainly with the analysis of states \cite{seqAn}. The co-presence network enables us to analyze groups of persons using network analysis methods such as cuts, cores, islands, network clustering and blockmodeling \cite{understand}.
The proposed approaches will be illustrated on the dataset Zvezoskop of Slovenian politicians \cite{ostro}.
\begin{thebibliography}{99.}%
\bibitem{understand} Batagelj, V., Doreian, P., Ferligoj, A. and Kejžar, N.: Understanding large temporal networks and spatial networks.
%: Exploration, pattern searching, visualization and network evolution.
(Vol. 2). John Wiley \& Sons (2014)
\bibitem{ostro} Oštro center: Zvezoskop. % -- an interactive visualization of career paths of active political officials. \\
\url{https://www.zvezoskop.si/en/} (accessed March 31, 2024)
\bibitem{seqAn} Ritschard, G., Studer, M.: Sequence analysis and related approaches: Innovative methods and applications. Springer Nature (2018)
\end{thebibliography}