How to learn enough data mining to be dangerous in 60 minutes

ABSTRACT

The field of data mining provides some methods highly relevant to researchers when mining software repositories. Whether one predicts bug locations, discovers hidden architectural structures and software patterns, or identifies experts of modules, data mining algorithms are usually the working horses for these studies. The goal of this tutorial is to convey some of the most relevant theoretical foundations and practical issues when using data mining algorithms.

The tutorial will first discuss the usual data mining tasks (prediction, filtering, smoothing, and elucidation of the most likely explanation or structure). Then, it will introduce a general framework for data mining paving the way to explain the functionality of some of the most used data mining algorithms. The tutorial will close with an overview over the typical evaluation methods for induced results and a number of pointers for further study. Where possible, it will use examples from software engineering.

INDEX TERMS

Primary Classification:
  H. Information Systems
  H.2 DATABASE MANAGEMENT
      H.2.8 Database applications
          Subjects: Data mining

Additional Classification:
  D. Software
  D.2 SOFTWARE ENGINEERING
      D.2.4 Software/Program Verification
          Subjects: Statistical methods

General Terms:
Measurement, Performance

Keywords:
mining software repositories