Citer Profiling Engine

Product Description

MindCite Profiler is a scalable , high-performance text profiling tool, capable of matching thousands of documents a minute against a large set of profiles.

Features:

• Supports up to 100,000 profiles per server.
• Keywords match the document text, as well as metadata.
• Profiles are Boolean expressions consisting of AND, OR and NOT operators. The expression syntax is structured, based on XML.
• The Boolean expression allows combination of data and metadata conditions in the same profile.
• The number of words used in a single profile can be up to 50.
• The engine supports the NEAR operator (directed) with distance designation.
• A pluggable mechanism allows the engine to integrate with various systems in order to notify the client of profile matching.
• The Engine invokes match notification which include:
  • The document identifier.
  • The matched profile ID.
  • The match score.
  • The words which caused the match.

Profiler

• The engine supports the following languages and features:

• Other languages can be added by MindCite upon request in a pluggable manner.
• The engine is able to extract text from the following file formats:
  • MS-Word
  • MS-Excel
  • MS-PowerPoint
  • PDF
  • HTML
  • XML
  • RTF (MS-Write / WordPad)
  • EML (RFC 822) including attachments.
• API:
  • The engine exposes API which enables feeding data (documents):
    By invoking an RPC call with the document stream
    By an RPC call designating the file location.
  • The engine exposes API which allows profile-management:
    Add, Remove and update profiles.
    Profile can be added dynamically while system is running.
• Optional (Not Included) features – to be developed upon need:
  • Wildcard support (*,?).
  • Soundex (phonic matching).
  • Support for WordPerfect® file-format.
• Engine is cross-platform and can be run on:
  • Microsoft Windows 2000 Server/2003 server.
  • Linux (kernels 2.4 and above).
  • Solaris (Solaris 10/ OpenSolaris recommended).