A total of forty-six causes of past landscape changes were
extracted from a content analysis of the interviews .We applied the
clustering procedure: agglomerative hierarchical clustering to
identify the causes of past landscape changes. The overall structure
of the clusters is over 0.5 (S ¼ 0.54), which shows than an acceptable
structure of the clusters was found. The optimal number of
clusters was five, as the location of the ‘elbow’ (the minimum
‘drop’) falls within this value (Fig. 5). The dendrogram was cut at 5
clusters (Fig. 6) and the silhouette plot showed that the causes of
landscape changes as they are perceived by stakeholders are
gathered within five large clusters, namely ‘increasing tourism,’
‘land tenure and social changes,’ ‘land use intensification,’ ‘postcommunist
transition’ and ‘foreign investments’.
A total of forty-six causes of past landscape changes wereextracted from a content analysis of the interviews .We applied theclustering procedure: agglomerative hierarchical clustering toidentify the causes of past landscape changes. The overall structureof the clusters is over 0.5 (S ¼ 0.54), which shows than an acceptablestructure of the clusters was found. The optimal number ofclusters was five, as the location of the ‘elbow’ (the minimum‘drop’) falls within this value (Fig. 5). The dendrogram was cut at 5clusters (Fig. 6) and the silhouette plot showed that the causes oflandscape changes as they are perceived by stakeholders aregathered within five large clusters, namely ‘increasing tourism,’‘land tenure and social changes,’ ‘land use intensification,’ ‘postcommunisttransition’ and ‘foreign investments’.
การแปล กรุณารอสักครู่..