Supplement to: Developing a Single Model and Test Prioritization Strategies for Event-Driven Software

This section presents detailed results for our study for RQ1 using the APFD metric. For each application, we divide the prioritization criteria into two sets (static and usage event-based) and present them in two graphs to better show the trends. The graphs are shown in log-scale. The controls (G-Best, Random, and G-Worst) appear in each graph. We also present the APFD values for the different prioritization criteria in tabular form. Due to space constraints, the tables report the APFD after every 10% increment of the test suite execution. In each table, we highlight the best APFD values for each increment with a bold font, along with the results for the G-Best control. TerpCalc: Figure 1 reports the APFD for the 10 prioritization criteria and 3 control criteria. This data is also shown in Table 10 in [1]. TerpPaint: Figure 2 and Table 1 show the results of our 10 prioritization criteria in relation to our controls of Random, G-Best, and G-Worst. In this study, PV-LtoS achieves the best rate of fault detection, followed by 2way. The Weighted-Freq and MFPS prioritization criteria are the 3 and 4 best, with Weighted-Freq providing