Technology

QUIID provide solutions for automatic knowledge discovery from relational databases. From your database(s) and according to your problematic, we help you to express what kind of information you need to discover, how it can be extracted and which views makes results analyses easier.

As we support you during your whole process, we are able to provide a tailored and turnkey solutions fitted to your needs. Our results are fully understandable and will create added value by highlighting new knowledge or leading you towards new R&D opportunities. Our technological core is a large library of data structures and statistics and data mining methods called KAD (Knowledge from Analyzed Data). The main task of KAD is to identify contrasting information inside a database: we deduce the set of descriptors (called pattern) which is best correlated to a subpopulation than to another one. We usually exploit a biological property activity to split the database (e.g. the toxicity of chemicals, the disease of patients…).

These patterns can be analyzed to deduce new knowledge or use in a larger process. In this case, discovering the patterns correlated to a property corresponds to the learning stage of the process: with these contrasting information, we could predict the property of new entities (e.g. the toxicity of chemicals, the diseases of patients, …) by seeking after learning patterns occurrences.

Characterization and prediction process is one of the several solutions provided by our technology. We support you in the workflow process design by picking modules to read, mine and view data and results. Each module is fully configurable to fit exactly your needs.

KAD is dependent of external open source tools:

Our software must be installed on a GNU/Linux server, we recommend the use of Debian in its stable version. The access to the graphical unit is done via the Mozilla Firefox browser. If you wish, you can visit the demo version of KAD.