Institutional-Repository, University of Moratuwa.  

Application of machine learning for extracting programming language constructs from 4GL legacy code

Show simple item record

dc.contributor.advisor Perera, AS
dc.contributor.author Subasinghe, WSAIlC
dc.date.accessioned 2017-01-31T08:55:28Z
dc.date.available 2017-01-31T08:55:28Z
dc.identifier.uri http://dl.lib.mrt.ac.lk/handle/123/12343
dc.description.abstract With the progression and innovations of the Information Technology industry, computer systems have become not only a part of an organization but the heart of it that drives their daily routines and manages and tracks the entire business process for most enterprises and for decades Advanced Business Languages (ABL) have been evolving to provide successful economic solutions to drive these businesses. Progress 4GL (Fourth Generation Language) is one such Advanced Business Language where organizations have developed entire business process on for 30 years. However, with the advancement of Free and Open Sourced Software providing business solutions, some organizations using these legacy systems are looking for means of migration. Even though proprietary service providers exists for the migration process, organizations with decades old data are reluctant to use them for both cost and security reasons. Yet, in house development is also costly since ABL experts are very few and would require much time and effort to complete the process. This research project is focused on a solution to develop such expert system that can interpret progress 4GL code to aid not only enterprises with migration but also engineers to learn and understand the language logic with ease. With the use of the Machine Learning technologies where research concerning modelling human thinking into machines are popular, this thesis provides a Proof of Concept for a methodology in which, an expert system can be created to read 4GL code, analyse the code, understand and infer the code logic and output the workflow in a graphical Flow Chart format. The prototype is run through several training 4GL programs to evaluate the implementation of the proposed theory. Current application proves to be successful for code with simple syntax and leaves room for further improvements to the system that can be enhanced to process 4GL’s many complex and evolving constructs and also the possibility of translating to a different language. en_US
dc.language.iso en en_US
dc.subject Expert Systems, Natural Language Processing, CLIPSJNI, Progress 4GL, mxGraph, Java-ML, Proparse en_US
dc.title Application of machine learning for extracting programming language constructs from 4GL legacy code en_US
dc.type Thesis-Abstract en_US
dc.identifier.faculty Engineering en_US
dc.identifier.degree M.Sc. en_US
dc.identifier.department Department of Computer Science & Engineering en_US
dc.date.accept 2015
dc.identifier.accno TH3101 en_US


Files in this item

This item appears in the following Collection(s)

Show simple item record