http://reposit.library.du.ac.bd:8080/xmlui/xmlui/handle/123456789/112 2026-04-28T09:52:19Z 2026-04-28T09:52:19Z Rahman, Md. Masudur http://reposit.library.du.ac.bd:8080/xmlui/xmlui/handle/123456789/4813 2026-04-13T04:33:05Z 2026-04-13T00:00:00Z

Minimizing Maintenance Cost through Prioritizing Refactoring of Code Smells Rahman, Md. Masudur Code smells are indicators of poor design practices that increase complexity, reduce comprehensibility, and hinder maintainability. While they do not directly affect system functionality, their long-term presence leads to higher maintenance costs and degraded software quality. Refactoring is the primary strategy to remove or minimize code smells. However, addressing all smell types is impractical due to time, budget, and resource constraints. Thus, identifying and prioritizing the most impactful code smells is essential for efficient maintenance and sustainable evolution of software systems. Code smell prioritization in existing literature is largely system-specific and often requires significant developer intervention. In other words, prioritization is typically applied to individual systems and tends to vary across different contexts. To address this research gap, the research conducts a large-scale empirical investigation aimed at establishing a generalized prioritization of code smells based on their impact on software quality and maintainability. Thirteen common smell types were examined across 35 open-source Java projects, analyzing their relationvi ships with 25 internal software quality metrics such as size, complexity, coupling, etc. as well as two maintainability metrics such as change-proneness and faultproneness. The study also incorporated perception-based insights from developers to compare subjective judgments with metric-driven analysis, uncovering notable discrepancies between these two. Finally, the research identified those code smells that exert the most detrimental effect on program comprehensibility, which is a key factor in reducing long-term maintenance costs. The results demonstrate that code smells vary in their degree of impact. Highpriority smells include Anti Singleton, Long Parameter List, Class Data Should Be Private, and Blob. Moderate-priority smells consist of Long Method, Complex Class, Large Class, Refused Parent Bequest, and Spaghetti Code. Finally, lowpriority smells are Speculative Generality, Many Field Attributes But Not Complex, Base Class Should Be Abstract, and Lazy Class. Alignment between developers’ perceptions and system analysis was observed for 61.54% of smell types, while 38.46% diverged, highlighting the need for developers to refine prioritization strategies to minimize maintenance costs. Finally, smells such as Long Method, Spaghetti Code, Refused Parent Bequest, and Anti Singleton were found to significantly degrade comprehensibility, with a correlation coefficient of −0.56 indicating that higher impact scores reduce comprehensibility. To summarize, the contributions of this research include an empirically derived prioritization of code smells, a comparative analysis of perception-based and metric-driven prioritization, and a synthesized dataset of 74,253 smelly files across 13 types. Collectively, these findings provide practical guidance for developers to focus refactoring on the most impactful smells, improve software quality, maintainability, and reduce long-term costs, while also supporting researchers in developing innovative refactoring tools and advancing future work in code smell management. This thesis is submitted for the degree of Doctor of Philosophy.

2026-04-13T00:00:00Z Hossain, Afzal http://reposit.library.du.ac.bd:8080/xmlui/xmlui/handle/123456789/4110 2025-04-20T05:05:42Z 2025-04-20T00:00:00Z

Ensemble Learning Algorithms for Classification Tasks in Natural Language Processing (NLP) Hossain, Afzal Natural Language Processing (NLP) encompasses a multitude of practical applications, including Information Retrieval, Information Extraction, Machine Translation, Text Simplification, Sentiment Analysis, Text Summarization, Spam Filtering, Auto-prediction, Auto-correction, Speech Recognition, Question Answering, and Natural Language Generation. Many of these applications are essentially classification tasks, which can be performed by machine learning models. Ensemble techniques within machine learning involve combining multiple models to improve predictive performance compared to individual models. This thesis explores the application of ensemble learning techniques to improve classification performance in NLP tasks. Various ensemble learning techniques, including bagging, boosting, random forest, and voting, are explored and experimented with. For each ensemble method, common base models, such as Support Vector Machines (SVM), Naive Bayes, Decision Trees, and K-Nearest Neighbor (KNN), are employed. Various evaluation metrics commonly used in NLP classification tasks are used, including accuracy, precision, recall, F1-score, and time complexity of the algorithms. The findings of the thesis suggest that ensemble methods, especially boosting, generally perform better than traditional machine learning methods for NLP classification tasks. The thesis also describes the modification of two ensemble models – firstly, majority voting is modified for the situation when a tie occurs, and secondly, bagging is modified with a different type of sampling. Both of these methods result in improved performances in the datasets. Overall, the research work provides a comprehensive overview of ensemble learning algorithms and their applications in improving classification performance in NLP tasks, backed by theoretical discussions, case studies, and experimental results. This thesis is submitted for the degree of Master of Philosophy.

2025-04-20T00:00:00Z Uddin, K. M. Aslam http://reposit.library.du.ac.bd:8080/xmlui/xmlui/handle/123456789/3438 2024-11-17T07:25:12Z 2024-11-17T00:00:00Z

A NOVEL BUG TRIAGING STRATEGY USING DEVELOPER RECOMMENDATION AND LOAD BALANCING MODEL Uddin, K. M. Aslam Bug triage is essential in efficiently assigning bugs to developers by leveraging past experiences. Without this crucial process, experienced developers may be inundated with assignments, while newer developers may be underutilized. Furthermore, improper bug distribution among different developer types can lead to various issues, including delays, errors, decreased capacity, and diminished job satisfaction. Previous bug triaging methods often do not account for newly joined developers, making them ineffective in recommending these developers for bug assignments. Consequently, these methods lead to improper task allocation, denying new team members valuable learning opportunities during bug resolution. Furthermore, prior research tends to overlook workload distribution among different developer categories, neglecting the need to balance bug assignments among experienced developers, newcomers, and those with varying skill levels. To address these issues, there is a need for an automated bug triaging technique that not only includes new developers but also prioritizes workload distribution among different developer categories. Therefore, this study introduces a novel bug triaging strategy that combines two pivotal models: Bug Solving Developer Recommendation Model (BSDRM) and Developer Scheduler (DevSched). The first model, known as the BSDRM, forms the core of automated bug triaging. BSDRM harnesses the power of Machine Learning (ML) algorithms and historical bug reports to intelligently suggest developers for specific bug resolution tasks. To achieve this, Eclipse, Mozilla, and NetBeans datasets are aggregated and split into training and testing sets. Subsequently, a sentence-embedded model is crafted from the training set, generating a developer-specific word repository. In contrast, the test set is transformed into a vocabulary list using an embedded model. BSDRM identifies eligible developers by matching their developer-specific word repository with the bug report vocabulary list via K-Nearest Neighbour (KNN) analysis. These developers are then categorized into three groups: experienced, newly experienced, and fresh graduate developers, utilizing a classification model comprising various ML algorithms Decision Tree (DT), Extra Tree (ET), AdaBoost (AdC), Bagging Classifier (BC), Gradient Boosting (GB), KNN, Nearest Centroid (NC), Bernoulli Na¨ıve Bayes (BNB), Multinomial Na¨ıve Bayes (MNB), Complement Na¨ıve iii Bayes (CoNB), Gaussian Na¨ıve Bayes (GNB), Logistic Regression (LR), Perceptron (Pr), and Multi-Layer Perceptron (MLP). Remarkably, the Bagging Classifier exhibits outstanding performance, achieving 96.59% accuracy in classifying developers with varying experience levels. In tandem with BSDRM, this study introduces the second model, DevSched, which assumes a critical role in balancing developer workloads. DevSched factors in workload distribution, developer proficiency, and bug characteristics. It generates multiple developer profiles based on historical bug reports and assigns bugs to developers by assessing the highest similarity between bug vectors and developer corpora. DevSched also dynamically adjusts developer workloads and refines their ratings based on performance. The study utilizes bug reports from Eclipse, Mozilla, and NetBeans to evaluate developer performance in the bug-triaging process. DevSched efficiently assigns and balances bugs among various developer categories, resulting in significantly reduced standard deviations for Eclipse, NetBeans, and Mozilla datasets compared to conventional bug distribution processes. This meticulous process is reiterated for each bug, ensuring optimal resource allocation and timely resolution of critical issues. The proposed study will collectively enhance bug resolution efficiency, optimize developer workloads, and ensure that both experienced and newer developers are judiciously utilized in the bug triaging process. A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of Master of Philosophy (MPhil).

2024-11-17T00:00:00Z Mahmud, Khaled http://reposit.library.du.ac.bd:8080/xmlui/xmlui/handle/123456789/3023 2024-02-18T07:01:01Z 2024-02-18T00:00:00Z

Sustainable Economic Growth for Developing Countries through Fintech Ecosystem: A Case Study on Bangladesh Mahmud, Khaled This study undertook a detailed investigation of financial technology (fintech) and its role in promoting sustainable economic growth for developing countries. In doing so, Bangladesh was the focal point. Through methods, both quantitative and qualitative this research aimed to answer four key questions related to fintech ecosystem readiness, customer readiness for fintech, fintech adoption factors, and fintech for sustainable economic development. The current study investigated the state of the fintech ecosystem in Bangladesh and its readiness with regard to ecosystem players. Narrative analysis, global comparison, and expert panel opinion point to a lack of fintech service and business model diversity in the still-developing fintech ecosystem of Bangladesh. It is found that Bangladesh is in the second stage of a three-step ecosystem development process. With a prudent, time-appropriate, and transparent policy framework, we believe that the fintech ecosystem in Bangladesh can contribute to sustainable development in the long-term. Apart from secondary datasets, this study conducts the National Citizen Survey (NCS) (N=1282). The nationwide representative sample was constructed through poverty-based stratified random sampling. Data were collected from 16 districts across Bangladesh. The NCS dataset provided the foundation for descriptive analysis and quantitative modeling. It incorporated demographic, economic, financial, technology usage, sentiment, and other variables related to fintech use. To the best of our knowledge, such a representative dataset on overall fintech use in Bangladesh is a first. This study also proposes the Customer Fintech Readiness (CFR) index to measure overall customer readiness for innovative fintech use. Given the hitherto absence of a measurement scheme for fintech readiness – as opposed to generic technology readiness, the CFR index considers seven key dimensions of customer readiness for fintech use and offers a customer fintech readiness measurement scheme. It has been found that Bangladesh is in the 26 th percentile of customer fintech readiness – lagging significantly behind in multiple dimensions e.g., financial conditions, existing fintech usage, etc. This study also deploys Recursive Feature Elimination (RFE) with multivariate logistic regression to model adoption factors of fintech. Among 133 features in the original model, 55 were preserved. Of these, 26 are found to be significant as determinants of fintech adoption. Importantly, 14 of these are related to customer concerns with various aspects of fintech use. Thus, customer concerns are major factors of fintech adoption in Bangladesh. Therefore, an effective way to raise adoption in the future is to address concerns and build customer trust. Finally, this study adopts the Case method, panel data regression, and univariate analyses with quartile-comparison to investigate the relationship between fintech and sustainable economic development. Across these three approaches, fintech’s contribution to sustainable economic development was evident. Particularly, our panel data model suggests that fintech channels like Automated Teller Machines (ATM) and debit card usage growth directly contribute to macro-level economic growth. More importantly, results from univariate analyses suggest that countries with higher growth in fintech channels e.g., debit card ownership, mobile money, digital payments, and wage distribution through cards also experienced higher growth in SDG index score, Goal 1, Goal 8, Goal 9, and Goal 11 scores during the period from 2014 to 2021. However, there were important nuances in results across these goals and between the two panels used: (a) all countries and (b) lower middle-income countries (LMIC) only. For LMIC, debit card ownership and digital payment showed the most significant association with progress in selected indicators. Further, univariate results point to a surprising lack of association of fintech with promotion of gender equality – thereby leading to further questions on effective ways to realize fintech’s transformative potential for women. We hope that the recommendations suggested in this report will contribute to the development of a more dynamic and vibrant ecosystem for sustainable economic growth in Bangladesh – and in developing countries across the world. A Dissertation Submitted to the Institute of Information Technology (IIT) University of Dhaka.

2024-02-18T00:00:00Z