Intelligent News Aggregator for German with Sentiment Analysis - Smart Information Systems: Computational Intelligence for Real-Life Applications - page 39

targets our future work will include on the one hand the extension of our corpus by
more text-anchored sentiment targets and on the other hand we will shift our work
toward topic-oriented sentiment target detection that aims at determining supporting
or opposing statements [ 34 ].
To extract even more opinionated passages from news articles in the future work
we not only want to consider quotations but also include other news article parts.
Furthermore, a context-dependent sentiment analysis requiring world knowledge or
interpretation could retrieve additional subjective text [ 27 ].

1.5 Application

We present the results of our news aggregator to users via a web interface. The
home page of our news'd demonstrator provides an overview of the currently most
important events discussed in the news. In order to tackle the enormous amount
of news material, our interface implements various ranking scores. Besides sorting
news clusters by actuality or size, users may navigate news clusters ordered by
their “hotness”. The hotness measure combines different cluster characteristics. It
weights appropriately a cluster's total growth since, its creation time, and its recent
growth in a sliding time window to calculate one score that indicates how “hot”
the news cluster is. Figure 1.7 shows the main page of our news'd demonstrator. As
in other commercial news portals our interface also allows browsing news events
by categories like “Politics”, “Economy”, etc. Each department page is organized

Fig. 1.7 The main page of the web application presenting the current top events