Capturing and analyzing streams of information
The internet provides constant, massive streams of information.
These streams grow and change whenever there is new information somewhere (news, alerts, blog entries, tweets, website changes, etc.).
Information comes to the user
Thanks to technologies such as web feeds, it is relatively easy to capture streams of information.
However, it is generally not possible to manually deal with all the information delivered by web feeds.
The amount of information is too great, and new information keeps coming up.
News, blog entries, alerts, web site changes, tweets, and mashups deliver thousands of potentially relevant pieces of information per day.
Enabling users to effectively deal with streams of information is the goal of mergeflow stream mining.
Analyzing streams of information
mergeflow stream mining enables users to grasp news without having to read them all in detail.
Our technology aims to address questions such as these:
How can I get a quick overview of current events?
Our stream mining technology clusters news by topic.
Thus, information on the same topic are in one place and do not have to be scrambled together manually.
Where do new topics come up?
"Frequent" does not necessarily mean "interesting".
On the contrary: many topics are particularly interesting as long as they are still brewing;
this implies that they are not (yet) frequent.
Our stream mining technologies help users detect such topics before they become part of the information mainstream.
How do topics develop?
mergeflow stream mining shows how topics develop over time.
Users can see how a topic spreads across different types of sources (e.g. mainstream vs. special interest),
and whether a topic relates to other topics (or whether it develops such relationships).
Put items of information into context
Interpreting information is all about context.
Context can drastically change how one and the same piece of information is interpreted.
Context can be defined by the information source, its thematic context, etc..
mergeflow stream mining helps generate context automatically.
For instance, we address the following questions:
What type of source does a piece of information come from?
Knowing more about the source of information often means knowing more about the information itself too.
Our technologies can automatically distinguish special interest sources from mainstream sources, for example.
We also automatically identify 'leading sources', i.e. sources that tend to report earlier than others on certain topics.
Where do events take place?
Whether it is the activities of a company, the spread of diseases, or economic topics (e.g. inflation, economic situation);
geographic context can play an important role in interpreting news.
Ouz technologies automatically detect locations associated with events. Thus, a user can get a geographical view on news and events.