|
| |
| Overview | New Indexation Techniques |
|
|
|
|
The majority of tools used for searching textual data rely on the Boolean model. Every item of text is associated in a matrix simply listing the terms and their position in the text. As a result the real importance of each term can be significantly overlooked or missed within the matrix, the sought after criteria being simply either « present » or « absent ». The expressions and specific terminology of a particular Organisation, whilst essential to the business, are rarely taken into account or are searched only as lines of characters without any special semantic value.
AMI’s indexing technology goes well beyond the simple Boolean index. Every document is subject to specific analysis that allows precise identification of the essential elements of the text from which AMI creates a Document Signature. In the same way you might mark important sections of a paper document with a highlighting pen or with different colours, AMI identifies and organises the terms of a document according to their level of importance.
As a result of the “Indexation by Signature” process (which can additionally be complimented by a Boolean index position) it is immediately possible to :
| • |
Return results of a very much higher relevance level. |
| |
| • |
Share the information: The content of a document can itself become the question allowing further discovery of all important associated information. |
| |
| • |
Organise a body of documents qualifying them by content: clusters, classifications, categories, etc. |
| |
| • |
Capture the information and, if need be, de-duplicate it according to the content. |
| |
| • |
Extract the key sentences of the text (quotations). |
Completely automatic in function (see the interview by Eric Fourboul, AMI’s Chief Technology Officer), the Document Signature technology is driven by AMI patented technology.
 Further information on Document Signature
|
|
|
