|
What is it?
Unlike keyword search systems that match exactly what you "say," concept search systems determine what you mean. Concept search engines aim to do what you naturally do in conversations with others, i.e., account for the many different ways people express the same idea.
Concept-based searching organizes unstructured information by mapping relationships between each word and every other word in large sets of documents. This association of words based on the context in which they are used creates a form of numeric "meaning" and "understanding," which enables users to quickly identify relevant documents and related concepts within vast amounts of data.
For
example, suppose someone sends you the following
email:
"Someone in the executive suite is going to
get canned, and it's not going to be pretty."
You can tell from the context that it doesn't mean someone is going to be preserved like a can of peaches and that your friend is not referencing beauty problems. To you, the meaning is quite clear. Similarly, if you were reviewing case documents and entered "fired from job" in a concept search engine, the application is smart enough to exclude information such as flames, smoke and fireplaces. The concept search engine will, however, effectively expand the search to dismissal, separation, layoff, suspension, etc.
Why is this so important for electronic
discovery?
Being able to find and review documents that are conceptually related to initial search queries, allows legal teams to thoroughly identify and review relevant and privileged documents more quickly and more accurately than with traditional search tools. Unlike keyword searching which requires skillful use of Boolean operators like "AND" and "OR," concept searching requires no formatting or syntax - reviewers can enter a natural language query or paste entire paragraphs from a relevant document, and within seconds receive a list of all related documents, ranked by relevancy.
Key Benefits of Concept Searching with Prevail:
- Quickly
find all relevant documents in your data set,
even those that don't share any words with your
initial search queries.
- Use unformatted,
natural language queries, eliminating the need
for managing complicated formatting or syntax.
- Efficiently
manage extremely large queries—including
entire documents—without performance
degradation. Scales seamlessly to handle large
(multi-terabyte), heterogeneous collections.
- "Concept space" is
built solely on your data, ensuring that search
results are truly relevant to your document
collection.
- Clear
highlighting of differences between keywords
and concepts.
|
