This project is a public archive of the practical project of Data Analysis in Software Engineering a curriculum unit of the Masters in Software Egineering at the Faculty of Enginnering of the University of Porto.
Our main goal is to predict if a Java class contains errors or not. We used a dataset from OpenStaticAnalyzer for java, on which each observation represents a Java class characterized by a set of quality metrics.
As of 2017, Software FAIL WATCH estimated that companies and organizations lost about 1,7 trillion USD (equivalent to 2% of world GDP). With an estimated number of persons affected of 3,7 billion (50% of the world population).
Finding the bugs in our projects has considerable advantages to our organization's quality standards and brings substantial economic benefits. We propose to develop a model that takes a set of statistical information and predicts if a Java Class will have bugs or not โ with a probability of 75% of finding bugs.