PREDICATED BUG SIGNATURE MINING VIA HIERARCHICAL INSTRUMENTATION METHOD DEVELOPMENT FOR BUG LOCALIZATION IN PROGRAM

• Main Author: Azka Widyanto, Muhammad
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/53923
• Institution: Institut Teknologi Bandung

Hiearchical Instrumentation Mining Predicated Bug Signature (HIMPS) is a method for localizing bug that produces bug signatures in the form of predicate set. In general, HIMPS consists of 2 stages, namely the coarse-grained level and the fine-grained level. These two stages have differences in the signature profile that is processed. At the coarse-grained level, input of a program is received to be used as a coarse-grained profile, so that a function list is generated with its Discriminative Significance value. Then, the function list is used as input in the initial stage of the fine-grained level which is then processed into a predicate set. This predicate set will be pre-processed and bug signature mining will be carried out so that the Discriminative Significance value that meets a threshold will be obtained for top-K bug Signature. This final project aims to test and analyze the performance of the HIMPS bug localization method against the Predicated Bug Signature. This is done by conducting theoretical studies as well as experiments using several input programs that have various number of predicates. With the various number of predicates, HIMPS performance is obtained which will be the data to be analyzed against the experimental results of the Predicated Bug Signature. In this final project, 2 programs were built, namely the Code Processor program and the HIMPS program. The experimental scheme is to run the two programs which are built to localize bugs for several input programs with a varying number of functions and predicates. From the experiment, the number of input program predicates, the execution time, and the amount of memory used during bug localization for an input program is obtained. After that, an analysis of the results of the HIMPS and Predicated Bug Signature experiments was carried out based on the number of input program predicates, execution time, and memory used. The experimental results show that HIMPS has a faster time performance than the Predicated Bug Signature. However, in terms of memory, HIMPS uses more memory than the Predicated Bug Signature.