Cortex Intelligence's text mining process is composed of five stages that were elaborated in detail in order to obtain the best results against the most frequent challenges faced in the text treatment market. The diagram that follows shows how these stages are chained:

The first step (1) is the collection of information and, to this end, the Cortex Intelligence robots navigate any environment to capure unstructured information, be it the internet or a company's internal data bases.

The following step (2) consists of pre-processing the collected texts. Intelligent agents process the text in order to extract and identify entities, adding meta-data to documents and enriching the information base.

Still in this second stage, these entities are connected amogst themselves through semantic relationships following a knowledge ontology according to Semantic Web (Web 3.0) patterns.

These factors confer the process a quality and reliability that is superior to any key-word based approach or purely statistical methods. The result was proved by academic studies undertaken by Cortex Intelligence while searching for the best approach for structuring texts.

The now structured text follows to the indexing stage (3), which is indipensable to the treatment of large volumes of data. Finally, the mining (4) itself applies high dimensionality statistical methods to each functionality demanded by the client.

Lastly, we have the participation of the user, who has the responsability of interpreting the results obtained, generating reports or starting new searches.

Differentiated Process, Better Results

The cortex Intelligence approach differs from the commonly used processes because it dedicates special attention to the pre-processing of the texts.

In general, text mining systems usually perform a simplistic pre-processing that often compromises the results of susequent stages. Cortex Intelligence, however, places a greater emphasis in the content of the texts because, as cited previously, it uses a multidisciplinary approach.

Another differentiating factor is the use of Dynamic Learning methods: the improvement of the system is continuous, because the algorithm accumulates the knowledge of previous runs.

An Intelligent Approach

Cortex Intelligence understands that text mining, beyond searching rapidly for the most relevant texts on a subject, should help in the arduous task of analyzing these texts. This could only be feasible by aggregating intelligence to the process.

Cortex's text mining intelligence is derived from the essence of Artificial Intelligence studies – the emulation of Reasoning.