Search engines work well if you are looking for specific and unambiguous information. But in order to find all relevant information about your local community, a political topic or a name shared by many, you need contextual information.
Suppose we want to get the best possible overview of local communities in Santa Cruz County, California. There are two Santa Cruz counties in the U.S., one in California and one in Arizona and there are multiple ‘Santa Cruz’es around the world, including brand names and people’s names.
Contextual analysis allows us to collect more relevant information. Relations to topics and geographical regions and similarities to other available information are all clues that can be used used to identify the news you are looking for.
This is how it works
Applied to finding relevant news, context analysis is a method used to analyze the environment in which the information is found.
When you select one or more general topics (A) and a geographical region (B), we can combine this information with the specific keywords you provide. Based on access to millions of articles, we then perform analysis that also takes into account language analysis such as the frequency of words, similarities between words and so on.
A keyword search becomes a search for matching patterns
The resulting pattern becomes a web of people, places, organizations and events centered around your topic of interest.
When new articles arrive, they are checked against this pattern and accepted as relevant information based on how well they fit in. The patterns gets adjusted as new information is added and consequently evolves over time.
Once an article is considered relevant, the final step is to find out how to combine relevance and importance with publishing policies that allows you to organize and prioritize information according to your preferences.