Intellexer Clusterizer

Intellexer Clusterizer enables effective organizing, normalizing, linking, and processing of documents, presenting of retrieval results in a more logical, structured, and searchable way. Meanwhile, it can provide information seekers with a means of browsing, and searching information efficiently in a user friendly format to fulfill the user’s information need.

Intellexer Clusterizer extracts terms (concepts and noun phrases) from unstructured text information, detects semantic relations (e.g., Verb-Subject-Object) between them, and finally forms clusters out of these terms, differentiating between broader and narrower meanings terms with the help of custom Intellexer Linguistic Processor and popular lexical database WordNet. Broader meaning terms indicate that the term represents a more inclusive concept, usually corresponding to a superclass link in an inheritance hierarchy. Narrower meaning terms indicate that the term represents a more specific concept. Particular consideration is given to the frequency of the term occurrence in the analyzed text.

Below is the example of a hierarchical tree of extracted terms.

Intellexer Clusterizer results can be successfully used for:

Business documents classifications;
Thesaurus management;
Semi-automatic ontologies creation;
Subject and domain information retrieval vocabularies;
Effective interchange with record management systems;

For developers and integrators

Use Case

Intellexer Clusterizer

Intellexer Clusterizer can be easily integrated into custom Document/Knowledge management systems using programming languages C/C++ and C#. Our SDK contains all necessary include files and import libraries for binding user applications with Clusterizer module.

Here is a C++ example of how to add Intellexer Clusterizer to your application:

								
#include <iostream>
#include <ClustyI.h>
#include <ClustyIInput.h>
#include <LPXml.h>

using std::cout;
using std::cerr;

using namespace NsSemSDK;

int main(int argc, char* argv[])
{
	try
	{
		char szPath[] = "../../LDB";                         //path to ldb
		char szQuery[] = "Electromagnetic radiation";        //query
		const int nRecordCount = 3;
		char szHeads[nRecordCount][100] =                    //heads
		{
			"Electromagnetic Spectrum - Introduction",
			"Electromagnetic Radiation",
			"Electromagnetic Waves"
		};
		char szBodies[nRecordCount][300] =                    //bodies
		{
		  "Each photon contains a certain amount (or bundle) of energy, and all electromagnetic radiation consists of these photons. The only difference between the various types of electromagnetic radiation is the amount of energy found in the ...",
		  "All these devices make use of electromagnetic waves. Radio waves, microwaves, visible light, and x rays are all examples of ... Scientists have observed that electromagnetic radiation has a dual \"personality.\" Besides acting like waves, ...",
		  "Light, microwaves, x-rays, and TV and radio transmissions are all kinds of electromagnetic waves. They are all the same kind of wavy disturbance that repeats itself over a distance called the wavelength. ..."
		};

		// provide path to license file
		SetClusterizerLicensePath("../../ISDK_License.xml");
		SetLPXMLLicensePath("../../ISDK_License.xml");
		
		// Load database
		CInterfacePtr<IClustyDB> pClusterizerDB(CreateClustyDB(szPath));
		CInterfacePtr<IRelTree> pRelTree;
		
		//Initialize IClustyInput interface
		CInterfacePtr<IClustyInput> pInput(pClusterizerDB->InputHA(""));
		
		//Set query
		pInput->SetQuery(szQuery, (int)strlen(szQuery));
		
		//Add records for processing
		for (int i = 0; i < nRecordCount; i++)
		{
			pInput->AddRecord(szHeads[i], (int)strlen(szHeads[i]), szBodies[i], (int)strlen(szBodies[i]), i);
		}
		
		//Notify end of input
		pInput->EndInput();
		
		//Get resulting tree
		pRelTree.Reset(pInput->CopyTree());
		
		//Print tree if not empty
		if (pRelTree->Size() != 0)
		{
			IRelTree::IIt* pIt = pRelTree->GetIt();
			do
			{
			   cout << pIt->GetText() << "\t" 
				<< (const_cast<IRelTree*>(pIt->GetTree()))->GetTreeID() << "\t"
				<< (unsigned)pIt->GetUniqueID(NULL) << "\n";
			}
			while(pIt->Next());
		}
	}
	catch (const CSemBaseException& x)
	{
		// Handle exceptions.
		cerr << x.what();
	}
	return 0;

As a result, all informative terms from sample documents will be grouped into hierarchical tree structure.

Intellexer Clusterizer

Overview