Technology: What makes
INSUMA search systems unique:
- Open architecture:
Modular system and open interfaces. Each module can be replaced or updated
independently from others - even at running systems.
- Neural nets and statistical features:
The latest scientific results in the area of neural nets and statistics make
our software "smart". These features are applied for
document classification, routing and clustering of document sets.
- Use of SOAP and XML/HTTP:
The search engine can be queried remotely via SOAP interface or XML queries
over HTTP. Proprietary databases or third-party software written in other
programming languages can be integrated compatible to licences in the same
- Intelligent crawler with rate functions: The crawler contains a filter
which allows Web documents to pass only if they are relevant (i.e., fit to
topics you defined previously). A rate function allows the user to mark only
relevant documents. This feedback is used to train the intelligent system.
- Existing module set: An existing vast module set allows to facilitate
search for the end user. Among others, we can provide support for German and
English morphology, phonetic search, typo correction and semantic search.
Support for many other languages (also cross-lingual search) can be offered
- Document filters: Document filters allow indexing of office documents
(DOC, RTF) as well as PDF documents in addition to the usual HTML and text
documents (Intranet search).
- Integration into frontends: A number of interfaces allow integration
to the search engine in arbitrary frontend applications (e.g., content management
systems (CMS), portal or authoring systems, Web frontends, etc.).
Modular structure of the Insuma platform. Proprietary modules can be
integrated via SOAP or XML/HTTP.
Programming languages used: C, C++, Python
Supported operating systems: Developed and tested under Linix distributions
SuSE, Debian and RedHat. Porting to Sun Solaris and Win32 are possible at