Skip to content

TsDD dataset

The TsDD dataset contains all experimental data used in the testability prediction and refactoring.

Machine learning testability prediction dataset available on Zenodo:

DOI

The current version of the testability prediction dataset (dataset06—version 0.6.x) contains the following files:

  1. DS060Raw.csv: Contains only source code metrics for 19,720 Java classes. The last column indicates the number of Java classes in the enclosing file of the presented Java class. Actually, SF110 contains more than 23K Java classes. We removed small projects and projects that most of their classes are data classes used as database models.