Because reading and analyzing a text is much more than just finding keywords out of a context that cannot be interpretated


  • Kernelis Descriptor is developed to automatically analyze texts and documents and thus prevents the interpretative bias of a human reader.

  • Kernelis Descriptor can also read and analyze documents faster and more accurately than a human reader would do.

  • Kernelis Descriptor can be used as a standalone software to analyze documents or can be added as a module to other software included in the Kernelis technology.


  • Kernelis-Descriptor analyzes documents accordingly to proprietary methodology in which semantic knowledge and advances in text mining are integrated. Hence, Kernelis-Descriptor is much more than a keywords finder.

  • Kernelis-Descriptor converts plain text into an alpha-numeric code that can further be interpreted and analyzed by the software and aligned with the sectoral requirements.

  • Kernelis-Descriptor is objective-oriented. This means that Kernelis-Descriptor is more efficient if the domain of activity is known and its jargon is integrated. 

  • Kernelis-Descriptor is compatible with non-alphabetic languages (Japanese, Chinese, Korean…).

The first step consists in the creation of topological mapping of the targeted domain of activity. This mapping results from the determination of weighed axes of activities sorted in U and D categories (approach consistent with the one developed in the context of Kernelis Resilience).

The topological map is then populated with bilingual semantic descriptors. According to the method, a same descriptor can be found in multiple locations of the map but its surrounding context (expressed as a distance to other descriptors) could then be different.

The plain text is analyzed by searching the semantic descriptors in its corpus. The text is then converted into a topological map and this map is compared to the sectoral map. Hence, the numerical code can estimate the relevance of the text with the addressed domain of activity.

“Kernelis Descriptor is much more than a simple statistic model that would only find, retrieve and index keywords from a formatted textual corpus.“


© 2019, by Kernelis.