Please use this identifier to cite or link to this item:
http://hdl.handle.net/10603/519196
Title: | Optimization of software maintenance Using hybrid deep learning neural Network based source code Summarization |
Researcher: | Chitti babu, K |
Guide(s): | Sethukarasi, T |
Keywords: | Computer Science Computer Science Information Systems Engineering and Technology neural Network software maintenance source code Summarization |
University: | Anna University |
Completed Date: | 2023 |
Abstract: | Software maintenance is the most critical and time-consuming newlineoperation performed in software industry, and as a result, it incurs significant newlinecosts. During software maintenance, developers devote a significant amount newlineof effort to understanding the source code. The Code Overview is highly newlinebeneficial for understanding code and looking for code since it generates brief newlinenatural language descriptions of the source code. The primary goal of this newlineresearch is to decrease the work required for software maintenance by newlinesummarising method modules in software artefacts. In this work using LSTM and a feature set, present a novel approach called CCDetector (Code Clone Detector) for detecting various forms of method level clones in software artefacts. The key contribution of newlinethis study is its four-fold LSTM-based technique to detecting various types of newlineclones. The first is structure-based detection, which compares the flow control newlinewithin the methods using the function flow graph, the second is logic-based newlinedetection, which compares the instruction to the flow chart symbols, the third newlineis flow-based detection, which compares the flow control within the methods newlineusing the function flow graph, and the fourth is I/O-based identification, newlinewhich verifies with a valid and invalid set of input-output pairs. With an newlineaccuracy rate of 0.99 and a recall rate of 0.98, CCDetector finds both newlinesyntactic and semantic code copies in the programme which is better than the newlinestate-of-the-art approaches. newline newline |
Pagination: | xiv,123p. |
URI: | http://hdl.handle.net/10603/519196 |
Appears in Departments: | Faculty of Information and Communication Engineering |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
01_title.pdf | Attached File | 2.58 MB | Adobe PDF | View/Open |
02_prelim pages.pdf | 2.2 MB | Adobe PDF | View/Open | |
03_content.pdf | 2.49 MB | Adobe PDF | View/Open | |
04_abstract.pdf | 2.49 MB | Adobe PDF | View/Open | |
05_chapter 1.pdf | 2.49 MB | Adobe PDF | View/Open | |
06_chapter 2.pdf | 2.49 MB | Adobe PDF | View/Open | |
07_chapter 3.pdf | 2.49 MB | Adobe PDF | View/Open | |
09_chapter 5.pdf | 2.49 MB | Adobe PDF | View/Open | |
10_chapter 6.pdf | 2.49 MB | Adobe PDF | View/Open | |
11_annexures.pdf | 124.29 kB | Adobe PDF | View/Open | |
80_recommendation.pdf | 86.09 kB | Adobe PDF | View/Open |
Items in Shodhganga are licensed under Creative Commons Licence Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).
Altmetric Badge: