In summary, to represent events in their geo-temporal context, most visualization systems either show the geographical distribution of the documents that discuss news events, or the information about the event itself.
We assume that a distribution design of the document tree is given, consisting of an XML kernel-document T[f1,…,fn] where some leaves are "docking points" for external resources providing XML subtrees (f1,…,fn, standing, e.g., for Web services or peers at remote locations).
In this article, we used the symmetrized Kullback Leibler divergences between the topic distributions of the documents.
Figures 4 and 5 show the distributions of the documents with regard to the numbers of sentences and event-predicates that they contain, respectively.
However, even after the calibration step, the color distribution of the scanned document may still significantly deviate from that of the original because of different photosensitive materials, refraction and reflection of different materials or the illumination conditions.
In this paper, the distribution of the publishing year, document types, language, subjects, journals, countries, institutions, times cited frequency of keywords, cluster analysis as well as collaboration of the WoS papers were thoroughly examined.
where, P(d i ) is the probability that a word will be observed in a given document d i, P(w j |z k ) is the probability of a particular word conditioned on latent topic variable z k, P z k |d i ) is the probability distribution of specific document over the latent variable space, and K is the number of topics.
Firstly, hot terms are extracted based on three features: the frequency of the terms used in the document collection; the location of the terms within a document; the breadth of terms distribution in the document collection.
