Prediction of Primate Splice Junction Gene Sequences with a Cooperative Knowledge Acquisition System

Wepropose a cooperative conceptual modelling environment in which two agents interact : the machineand the humanexpert. Theformer is able to extract knowledge from data using a symbolicnumericmachinelearning system, and the latter is able to control the learning processby acceptingand validating the machineresults, or by criticizing those results or the explanationthat the systemproduceson them. Theimprovmentof the conceptual modelling relies onthe cooperationbetweenthe twoagents. Results obtained withour methodon prediction of primatesplice junctionssites in geneticsequencesare far better than theses reportedin the literature with other symbolicmachinelearning systems, and are as better as theses obtainedwithsomeartificial neural networks methods reported at present. But in opposite to neural networks which lack of argumentation, our system provides the user a plausible explanationof its prediction.

• Type: Conference
• Year: 1993
• Where: ISMB
• Authors: Engelbert Mephu Nguifo, Jean Sallantin
• Keywords: Computational Biology, Conceptual Modelling, Conceptual Modelling Environment, ISMB 1993, Neural Networks