next up previous
Next: Dissemination via Web Server Up: Project Previous: Knowledge mining

Rule evaluation

The following questions will be posed to assess the quality of topological rules derived. i) Have new principles and common substructures been derived? ii) How readily comprehensible to an expert in protein structure are the rules? iii) How much of the data has been explained and how accurate are the rules (quantified by compression, coverage and accuracy). iv) Could these new rules be found by simpler approaches and if so how easily? v) Can these rules be used to improve secondary and tertiary structure predictions?



Marcel Turcotte
1999-10-20