Page ThesisSCFGs

((short description what this page is about))

(1) The software part and what has been done so far: - parsing of scfg data from files - transformation of arbitrary context-free grammars into Chomsky normal form - implementation of inside, outside and CYK algorithms (with data generation for backtracking) - unit-testing of the implemented parts

(2) Functionality yet to be implemented: - training from sequences (esp the inside-outside EM maximization step) - Covariance Models, if time permits. (Requirement for grades better than 'good') - unit test for added parts

(3) Application of the software - selection of training RNA set - training of a model - use some kind of discriminator(?) and see if we find anything (cross validation of training set?)