Optimize edilmiş makine öğrenim tekniklerine dayalı yazılım kusurlarını öngörmek için yeni bir yöntem

Loading...
Thumbnail Image

Date

2022

Authors

Yazıcı, Ali

Journal Title

Journal ISSN

Volume Title

Publisher

Open Access Color

OpenAIRE Downloads

OpenAIRE Views

Research Projects

Organizational Units

Organizational Unit
Software Engineering
(2005)
Department of Software Engineering was founded in 2005 as the first department in Ankara in Software Engineering. The recent developments in current technologies such as Artificial Intelligence, Machine Learning, Big Data, and Blockchains, have placed Software Engineering among the top professions of today, and the future. The academic and research activities in the department are pursued with qualified faculty at Undergraduate, Graduate and Doctorate Degree levels. Our University is one of the two universities offering a Doctorate-level program in this field. In addition to focusing on the basic phases of software (analysis, design, development, testing) and relevant methodologies in detail, our department offers education in various areas of expertise, such as Object-oriented Analysis and Design, Human-Computer Interaction, Software Quality Assurance, Software Requirement Engineering, Software Design and Architecture, Software Project Management, Software Testing and Model-Driven Software Development. The curriculum of our Department is catered to graduate individuals who are prepared to take part in any phase of software development of large-scale software in line with the requirements of the software sector. Department of Software Engineering is accredited by MÜDEK (Association for Evaluation and Accreditation of Engineering Programs) until September 30th, 2021, and has been granted the EUR-ACE label that is valid in Europe. This label provides our graduates with a vital head-start to be admitted to graduate-level programs, and into working environments in European Union countries. The Big Data and Cloud Computing Laboratory, as well as MobiLab where mobile applications are developed, SimLAB, the simulation laboratory for Medical Computing, and software education laboratories of the department are equipped with various software tools and hardware to enable our students to use state-of-the-art software technologies. Our graduates are employed in software and R&D companies (Technoparks), national/international institutions developing or utilizing software technologies (such as banks, healthcare institutions, the Information Technologies departments of private and public institutions, telecommunication companies, TÜİK, SPK, BDDK, EPDK, RK, or universities), and research institutions such TÜBİTAK.

Journal Issue

Abstract

Bu tezde, tüm gerçekleri motivasyon olarak kabul ederek yazılım kusur tahmini için yeni ve sağlam bir buluşsal güdümlü nöro-bilgisayar modeli geliştirilmiştir. Diğer klasik makine öğrenimi modellerinden farklı olarak, nöro-bilgisayar, özellikle Levenberg Marquardt Sinir Ağı (LM-YSA), doğrusal olmayan özellik öğrenimi ve dolayısıyla hatalı veriler için hayati önem taşıyabilecek uyarlamalı öğrenme açısından daha sağlam olarak kabul edilimektedir. Ancak, diğer makine öğrenimi modellerinde olduğu gibi, 17 giriş özelliği olanlarda da aşırı yüksek ağırlık tahmini nedeniyle yerel minimum ve yakınsama olasılığından kaçınılamamıştır. Bu gerçeği göz önünde bulundurarak, bu araştırma, öğrenme sırasında uyarlanabilir ağırlık tahmini ve güncelleme için YSA'ya yardımcı olamak amacıyla buluşsal model denilen yeni bir geliştirilmiş genetik algoritm sunark katkıda bulunmuştur. Burada buluşsal modelin temel amacı, LM-YSA'nın herhangi bir yerel minimum ve yakınsama sorunu yaşamadan üstün ağırlık tahmini, güncelleme ve dolayısıyla öğrenme elde etmesine yardımcı olmaktır. Sonuç olarak , önerilen nöro-bilgisayar modelinin hedeflenen yazılım hatası veri kümeleri üzerinde klasik sinir ağından daha yüksek doğruluk elde etmesine yardımcı olmuştur. Sınıflandırıcı veya makine öğrenimi iyileştirmesine ek olarak, bu araştırmada, herhangi bir sınıf dengesizliği, aşırı uydurma ve yakınsama olasılığının hafifletilmesine yardımcı olan özellik mühendisliğine de odaklanılmıştır.
In this thesis a novel and robust heuristic driven neuro-computing model was developed for software defect prediction. Unlike other classical machine learning models, neuro-computing, especially Levenberg Marquardt Neural Network (LMANN), is considered to be more robust in terms of adaptive learning, which can be vital towards non-linear feature learning and hence defect data. However, similar to the other machine learning models, the likelihood of local minima and convergence could not be avoided due to exceedingly high weight estimation for 17 input features. Considering this fact, this research contributed a novel improved genetic algorithm, say heuristic model was developed to assist ANN for adaptive weight estimation and update during learning. Here, the key purpose of heuristic model was to help LM-ANN gaining superior weight estimation, update and hence learning without undergoing any local minima and convergence problem. This as a result helped the proposed neurocomputing model to achieve higher accuracy than the classical neural network over targeted software fault datasets. In addition to the classifier or machine learning improvement, in this research the focus was made on feature engineering as well that helped alleviating any probability of class imbalance, over-fitting and convergence.

Description

Keywords

Bilgisayar Mühendisliği Bilimleri-Bilgisayar ve Kontrol, Genetik algoritmalar, Computer Engineering and Computer Science and Control, Kusur tespiti, Genetic algorithms, Nöromorfik, Flaw detection, Neuromorphic, Yapay sinir ağları, Artificial neural networks

Turkish CoHE Thesis Center URL

Fields of Science

Citation

WoS Q

Scopus Q

Source

Volume

Issue

Start Page

0

End Page

97