Applying data mining to contract inference
Nikolay Kazmin · Repository for Publications and Research Data (ETH Zurich) · 2010
This master thesis report will introduce the techniques that we use to automatically infer contracts for a given class.The report will cover briefly the building of the automatic test suite, the generation of the change profile and then will get into great details on the implication inference.It will focus on the decision tree and linear regression learning techniques to infer contracts.Experiments show that for standard data structure classes such as linked list, array, stack, query and hash table, 75% of the complete contracts can be inferred fully automatically.8 Conclusions 8.1 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .