Tecnicas de inferencia gramatical y aplicacion al procesamiento de biosecuencias (participacion desde colombia en el tercer año de este proyecto).

This project belongs to the artificial Intelligence area, more exactly to the automatic learning field oriented to the prediction and analysis of meaningful features in biosequences, that is, in DNa, RNa and protein chains. The automatic learning techniques object of this research, use the inductive inference paradigm from the language theory point of view, this is known as the grammatical inference area. This project has as its main goal to develop new grammatical inference algorithms to some well known language classes as regulars and context free languages. Correctness of algorithm proposed will be proved formally showing limit identification and empirically by experiments on synthetic corpora. Corpora will be available for other researches interested on them. Second goal is to apply grammatical inference to problems in bioinformatics area, more exactly to the prediction and analysis of features such as protein motives or coding regions on DNa and RNa chains. Technological transference would be achieved by means of development of software tools in Java language and a web interactive portal publishing those tools. Notice that grammatical inference produces true algorithmic (no heuristic) solutions and they use to have well funded formal properties.