Title :
Mining Attribute Lifecycle to Predict Faults and Incompleteness in Database Applications
Author :
Kaiping Liu ; Hee Beng Kuan Tan
Author_Institution :
Nanyang Technol. Univ., Singapore, Singapore
Abstract :
In a database application, for each attribute, a value is created initially via insertion. Then, the value can be referenced or updated via selection and updating respectively. Eventually, when the record is deleted, the values of the attributes are also deleted. These occurrences of events are associated with the states to constitute the attribute lifecycle. Our empirical studies discover that faults and incompleteness in database applications are highly associated with the attribute lifecycle. Consequently, we propose a novel approach to automatically extract the attribute lifecycle out of a database application from its source code through inter-procedural static program analysis. Data mining methods are applied to predict faults and incompleteness in database applications. Experiments on PHP systems give evidence to support applicability and accuracy of the proposed method.
Keywords :
data mining; database management systems; program diagnostics; PHP systems; attribute lifecycle; attribute lifecycle mining; data mining methods; database applications; fault prediction; incompleteness prediction; inter-procedural static program analysis; Data mining; Databases; Educational institutions; Predictive models; Support vector machines; Training; Vectors; Fault prediction; attribute lifecycle; data mining; incompleteness prediction;
Conference_Titel :
Software Engineering Conference (APSEC), 2013 20th Asia-Pacific
Conference_Location :
Bangkok
Print_ISBN :
978-1-4799-2143-0
DOI :
10.1109/APSEC.2013.39