Underneath the hood of the Imagisoft Ontology Engine

Recall that the Imagisoft semantic engien performs the following:

G3 = convolve(G1,G2)

The G1,G2,G3 are ontology maps in nested form. For each of the G, any node is the triplet:

(keyword,operator,link)

(Note: The Imagisoft engine allows proprietary extensions to the record set)

The convolution operation is a simultaneous traversal of both G1 and G2 together, producing G3 as a trace of the common path elements. At the core of the convolution operation are two operations:

  match(G2.node.op,G1.node.keyword) and match(G1.node.op,G2.node.keyword)

The first match advances along G2, the second along G1. When both matches occur, then another nested link is pushed onto G3. So G3 is the ongoing match paths along which G1 and G2 found agreement. A real world example of G3 is the search page results we get from searching he public web, and the G1 would be our search list. All the terminals of G2, the web ontology, terminate in URL and RDF resources.

There is no schema in the system, rather canonical ontologies are built up the oorporate mesh, using text search term and click thrus to mediate between authors and readers.

In the Imagisoft system there is a native semantic analyzer, programmed as a nested graph that moves raw text through the dictionary/reduction process to produce dense ontologies of author text, which is then dumped into the corporate mesh. So operation of the Imagisoft system is fairly easy, just blog some text and include any links. Then submit your post URL to the engine.

How can I get the engine!, you ask.
It will be available for download at source forge at some time in the future. That is how we do it here at Imagisoft.