In this paper, we attempt to extract important sightseeing spots so as to improve the convenience of tourists. To this end, we propose two measures, one is detour centrality which is a easiness measure of brief detour and the other is convenience centrality which is a accessibility measure, based on traditional closeness and betweenness centrality respectively. These traditional centrality measures do not intend for actual distances as weights on links therefore, our proposed measures are designed with distances between spots. Furthermore, by assuming that these natures like easiness of detour and accessibility of spots behave collaboratively, we also propose extended measures for an arbitrary set of spots. In our experiments using two real sightseeing spots datasets, we show that our extended method can extract an appropriate set of spots in terms of easiness of detour and accessibility, and these pivots extracted by two methods are substantially different.
抄録全体を表示