Towards Succinctness in Specification Mining: The Scenario-Based Case

Supporting Materials

David Lo (Singapore Management University) and Shahar Maoz (RWTH Aachen University, Germany)

Abstract

Specification mining methods are used to extract candidate specifications from system execution traces. A major challenge for specification mining is succinctness. That is, in addition to the soundness, completeness, and scalable performance of the specification mining method, one is interested in producing a succinct result, which conveys a lot of information about the system under investigation but uses a short, machine and human-readable representation.

In this paper we address the succinctness challenge in the context of scenario-based specification mining, whose target formalism is live sequence charts (LSC), an expressive extension of classical sequence diagrams. We do this by adapting three classical notions: a definition of an equivalence relation over LSCs, a definition of a redundancy and inclusion relation based on isomorphic embeddings among LSCs, and a delta-discriminative measure based on an information gain metric on a sorted set of LSCs. These are applied on top of the commonly used statistical metrics of support and confidence.

A number of case studies show the utility of our approach towards succinct mined specifications.

Case Study

Machine Settings:
              2Ghz Intel Core 2 Duo with 2GB RAM Fujitsu Laptop

    Windows Vista Business
              Mining program is written using C#.Net 2.0 compiled using VS.Net 2005

Target Program:

              crossFTP -- a commercial open source FTP server built on top of Apache FTP server.
 

List of Method Calls Traced:

              List of methods traced

 

Input:

  Trace Set (1876 events)

 

Output:

             All Succinct Scenarios Mined  (Significant, Non-Redundant, Delta-Discriminative) (11 scenarios)

 

Comparative Output:

             All Significant Symbolic Scenarios Mined  (118,012 scenarios)