Bachelor of Computer Science and Engineering (CSE)
Permanent URI for this collection
Browse
Recent Submissions
Item A Semantic Web Approach: Creating A Tourism Ontology For Chittagong District(International Islamic University Chittagong, 2023-07) Ullah, Md. Rohmat; Riyad, Suliman Hossain; Rahat, Rakib HasanThe purpose of this thesis is to utilize the Semantic Web, a web of interconnected mean ing, to develop a comprehensive tourism ontology for Chittagong district. This paper implemented ontology on tourism domain, proposed a general framework for tourism on tology and explained searching mechanism through Chittagong district tourism. Also, it presents different ways of reasoning the ontology. In general, ontology classifies the vari ables in need for some computations and creates interrelationships between them. The introduction of semantic web poses the demands for creating ontology in many domains. We have found that the utilization of ontologies within the tourism domain remains relatively limited. Notably, research into ontologies specifically focused on Chittagong tourism is entirely absent. To address this gap, this study proposed the development of a dedicated Chittagong tourism ontology. In the digital age, tourism thrives on readily available information and efficient organization. This thesis delves into the creation of a robust tourism ontology for Chittagong district, Bangladesh. By formalizing knowledge about tourism resources, attractions, and experiences, we aim to enhance information re trieval, facilitate data integration within the tourism sector. This ontology leverages the expressiveness of Web Ontology Language (OWL) to model the intricate relationships between various tourism entities. We capture details about historical sites, cultural at tractions, natural wonders, transportation networks, foods and accommodation options. The ontology also incorporates relevant concepts like accessibility, sustainability, and cultural sensitivityItem Approach to Improving Machine Learning Models for Intrusion Detection System(International Islamic University Chittagong, 2022-07) Labib, Ahmad Ibtisam; Chy, Shamsuddin Ahmmed; Hossain, Md. ShahriarIn today's digital world, there are several security risks that digital assets must contend with. Systems for detecting intrusions (IDS) are essential security tools that protect digital assets. But their usefulness depends on meeting strict accuracy requirements, and their effectiveness depends on timely alarms. This study offers a novel IDS model that combines deep learning and machine learning methods as a solution to these problems. The study applies several classification techniques, such as Gaussian Naive Bayes (GNB), Random Forest (RF), Decision Tree, K-Nearest Neighbors (KNN), Soft Voting, and Hard Voting, using the well known KDD Cup-1999 dataset. After a large-scale dataset was processed, the Decision Tree method performed better than the others, with a 99.9% accuracy rate. This study aims to investigate the effects of soft voting and hard voting, a novel application in IDS. Decision Tree proved to be the better performance in spite of these efforts. By offering information about algorithmic efficacy, the research advances the field of intrusion detection and helps decision-makers in the design and deployment of intrusion detection systems. These findings have implications for improving digital asset protection against changing cyber threats.Item ScrapVault – Which is a digital marketplace dedicated to buying and selling scrap materials.(International Islamic University Chittagong, 2022-06) Chowdhury, Md. Ashfiqul Alam; Chowdhury, Alim Ullah; Amin Joy, ShahriarScrapVault is an innovative digital marketplace that aims to revolutionize the scrap trading industry by providing a dynamic platform to connect environmentally conscious individuals and businesses. The website has been designed with a user-centric approach, offering an intuitive interface for easy navigation and efficient discovery of a wide range of scrap materials, including metals, plastics, and paper. The platform prioritizes security, offering a robust transactional framework to ensure peace of mind. Advanced geo-location features are integrated to help users locate nearby sources of scrap materials. ScrapVault is unique in its commitment to transparency, providing environmental impact metrics that allow users to track and quantify their contributions to sustainability. In addition to its functional aspects, ScrapVault also focuses on building a community of like-minded individuals and businesses. This community not only supports recycling efforts but also cultivates a sense of responsibility towards the environment. Overall, ScrapVault is a commendable project that contributes to the digital trans formation of the scrap trading industry. By balancing economic considerations with a commitment to sustainability, community engagement, and environmental responsi bility, ScrapVault has emerged as a trailblazing initiative towards a greener and more interconnected futureItem A Web Based Application For Home Service System(International Islamic University Chittagong, 2022-07) Obyedullah, A.S.M.; Shanto, Naimul Islam; Afridi, ShahidIntroducing a revolutionary home service system, an innovative web-based application poised to revolutionize the way service seekers connect with providers. This cutting-edge platform not only facilitates seamless discussions about diverse home services but also streamlines the entire booking process, enabling hassle-free online payments for the discerning user. In times of urgency, our system excels at handling quick service requests, sparing you the inconvenience of complicated bookings. Moreover, it opens up exciting job opportunities tailored to the specific needs of service providers. Picture this: users empowered to select their desired service, effortlessly logging into their personalized portals to manage their profiles directly from our website. With the seamless integration of computer and smartphone accessibility, connecting with service providers has never been more intuitive. Embrace the future of home services with our state-of-the-art platformItem An E-commerce recommendation system based on LightGBM Machine Learning Algorithm(International Islamic University Chittagong, 2022-07) Hera, Mohammed Ashrafujjaman; Chowdhury, Md.Injamamul Hoque; Tohidu, MohammadIn the dynamic realm of e-commerce, recommendation systems play a pivotal role in shaping user experiences and fueling business growth. This study advocates for a novel approach to online shopping recommendations, leveraging the power of the LightGBM machine learning algorithm. By focusing on item-to-item recommendations, our method ology seeks to elevate user satisfaction by swiftly and precisely offering customers highly personalized product choices. At the core of our recommendation system lies the fusion of item-to-item association analysis and user interactions, culminating in the delivery of accurate, real-time rec ommendations. This research contributes significantly to the e-commerce landscape by presenting a practical and scalable method that enriches customer experiences, conse quently amplifying sales and fostering customer loyalty. Through extensive testing and evaluation, our results underscore the transformative potential of the proposed item-to-item e-commerce recommendation system. This inno vative system stands poised to revolutionize digital commerce by providing users with pinpoint-accurate product recommendations. The seamless integration of machine learn ing, coupled with a focus on item-to-item relationships, not only expedites the decision making process for consumers but also cultivates a deeper connection between customers and the digital marketplace. In summary, our study demonstrates the capability of our approach to usher in a new era of precision and effectiveness in digital commerce, promis ing a paradigm shift in the way users discover and engage with products onlinItem Resiroute – A Platform For Finding Accommodations, Including Hotels, Houses, Apartments, And Shared Apartments(International Islamic University Chittagong, 2022-07) Uddin, Md Mosleh; Hamid, Syed Md Abdul; Sakib, Mohammed ArifullahThe travel industry is one of the fastest-growing sectors worldwide, with millions of people traveling for business or leisure every year. As more people seek personalized and unique travel experiences, there is a growing demand for alternative accommoda tion options that provide a more authentic and immersive experience. ResiRoute is a novel platform that aims to connect travelers with hosts offering unique and personalized accommodations, departing from the conventional hospitality models offered by hotels and resorts. ResiRoute is committed to prioritizing user experience and community building, redefining the landscape of short-term lodging through a fresh perspective on travel accommodation. The platform’s technical architecture includes robust user authentication, secure payment systems, recommendation algorithms, and a feedback mechanism, creating a seamless and efficient booking experience for users. The devel opment process of ResiRoute was comprehensive, focusing on creating a platform that provides a unique and innovative approach to the travel accommodation industry. The platform’s security measures ensure the safety and security of its users through strict verification processes and privacy protocols. Additionally, ResiRoute aims to contribute to sustainable tourism practices by promoting responsible travel and supporting local communities. The platform’s community-building approach fosters a shared experience between hosts and travelers, encouraging users to develop meaningful connections and contribute to local economies. By providing hosts with an additional source of income, ResiRoute offers opportunities for entrepreneurship and fosters economic growth in local communities. In conclusion, ResiRoute offers a fresh perspective on the travel accommo dation industry, providing a platform that connects travelers with hosts and promotes a shared community experience. The platform’s commitment to prioritizing user ex perience, community-building, and responsible travel makes it a unique and innovative option for travelers seeking a personalized and authentic travel experience. As the travel industry continues to grow, ResiRoute has the potential to revolutionize the way peo ple travel, contributing to cultural exchange, economic growth, and sustainable tourism practicesItem Malware Detection System using Machine Learning & Deep Learning Technique(International Islamic University Chittagong, 2023-07) Karim, Miskatul; Hossain, Md JubairThese days, with the amount and variety of malware increasing at an exponential rate, it is imperative to use innovative methods for quickly and precisely detecting this dangerous software. The fast velocity at which malware spreads makes manual heuristic inspections of malware analyses inefficient for both discovering new malware and keeping up with it. Machine-learning techniques have therefore become increasingly popular. By automating the examination of static and dynamic studies, these methods can classify unknown malware according to how close it is to recognized families and combine malware that exhibits similar behavior. While data mining and machine learning approaches have been applied in the past, this paper demonstrates how deep learning networks can improve accuracy even further. By building neural networks that have more potentially varied.Item Book Valley [An Online Web-Based Book Publication & Selling System](International Islamic University Chittagong, 2022-07) Chowdhury, Md. Hayat Hossain; Hoque, Md. Sohidul; NadimNowadays most of the work, entertainment facilities, and education is done by digital tools. Now we are living in a digital world where time is so valuable, that people want to lead hassle-free life. For a hassle-free centralized Online Bookselling flat form (Book Valley) for bookworm people and businessmen. This system gathers readers, writers, and publishers in one flat form. Readers can find a huge amount of books from anywhere on the earth and carry all the books with them all the time wherever they are. The E-book system is a blessing for book reader because it saves their time, carrying hassle, and storage costs. A writer can contact and send their creation to any publication over the world. Mostly new writers can easily reach the publisher for their desired publication. Book Valley provides e-books for readers for that reason the cost of publication will decrease and publishers can easily publish their books and sell them in online stores. By utilizing Book Valley’s features as a centralized system we can get a hassle-free book ecosystem where readers, writers, and publishers will benefit equally.Item QUIZ CRAFTERS - A Web-based Quiz Platform(International Islamic University Chittagong, 2022-07) Imon, Rakibul Hasan; Tazwer, Syed Fahim; Rahman, Koushik JamilurAmid the diverse array of quiz platforms, a pressing need for a universally accessible and free alternative has become increasingly evident, as prevailing solutions often come burdened with financial constraints. In response to this challenge, Quiz Crafters, our visionary project, endeavors to democratize educational technology by furnishing educators with a cost- effective and feature-rich alternative. Motivated by the limitations experienced with paid platforms like Socrative and driven by a passion to elevate user experience, our platform pledges user-friendliness, real-time engagement, data-driven insights, and customizability— all without imposing financial burdens on educators. In a landscape where the demand for accessible and budget-friendly interactive assessment tools in education is on the rise, Quiz Crafters emerges as a liberating alternative to existing platforms. Acknowledging the restrictions inherent in paid services, particularly exemplified by platforms like Socrative, our mission is to democratize educational technology. In commitment to removing barriers, our quiz taking website is meticulously crafted to cater to the diverse needs of educators, seamlessly accommodating both intimate classroom settings and expansive lecture halls. Through a dedicated emphasis on inclusivity and innovation, Quiz Crafters aspires to redefine the very landscape of interactive educational assessment. Our aim is to provide a transformative and easily accessible tool, not only for educators but for students as well. This project is a testament to our unwavering dedication to innovation, inclusivity, and the continuous advancement of interactive educational assessment tools.Item Brain Tumor Classification using Semi Supervised Self-Training Algorithm(International Islamic University Chittagong, 2022-12) Islam, Mohammad Nafizul; Ali, Mohammed Mushfiq; Abrar, Md ShafiulBrain tumors have been considered the world’s most dangerous medical condition. Worldwide, approximately 0.25 million people die every year due to CNS tumors and primary cancerous brains. The existing method is obtrusive, tedious, and sensitive to errors by individuals. To overcome the pitfalls mentioned above for brain tumor multi class classification, an approach named confidence regularized self training for brain tumor classification (CRST-BTC) has been proposed for the early detection of brain tumors in an efficient way. For the purpose of early diagnosis, this research introduces a novel approach to enhance the accuracy of brain tumor classification using confidence regularization self-training. The study systematically investigates the impact of varying percentages of unlabeled data and explores different backbone networks, with VGG16 emerging as the standout performer. Additional improvements, including integrating a dense layer and optimization algorithms such as Adam, contribute to superior clas sification accuracy. Surprisingly, the model performs better with 20% labeled data, challenging conventional expectations. The research leverages the SHAP framework for visualization, providing insights into feature importance. Benchmarking against state of-the-art models showcases the competitive edge of the proposed methodology. The VGG16 backbone with the Adam optimizer stands out, achieving 97% accuracy, 98.67% precision, 97.45% recall, and 98.01% F1-score. This study advances the field and pro vides valuable insights for practitioners in medical image analysisItem Digital Dining- From where you can order your desired food(International Islamic University Chittagong, 2022-07) Sazid, Mohammed Saiful Azam”Digital Dining” introduces a user-centric website focusing on streamlined food ordering and menu exploration for a fictional restaurant. This platform enables customers to seamlessly browse the menu, place food orders, and enjoy a hassle-free digital dining experience. The website prioritizes simplicity and convenience, allowing patrons to peruse the restau rant’s offerings and place orders effortlessly. While reservations aren’t facilitated, the emphasis remains on providing a user-friendly interface solely dedicated to exploring the menu and initiating food orders. By concentrating on efficient food selection and ordering processes, Digital Dining aligns with modern consumer preferences, catering to their desire for simplicity and ease in on line dining experiences.Item Dengue detection using Machine Learning Algorithms(International Islamic University Chittagong, 2022-07) Turhan, Nazmus Sakib; Hossain, Md. Hasibul; Hossain, Md. ReyadThe dengue virus, a global health concern with an estimated 400 million annual infec tions, has posed a critical challenge worldwide. Between January 1 and August 7, 2023, Bangladesh experienced a concerning surge in dengue cases, reporting over 69,000 con firmed cases and 327 deaths according to the Ministry of Health and Family Welfare. Timely and accurate diagnosis is necessary for intervention and treatment optimization, possibly relieving severe complications and saving lives. In response, this paper intro duces a machine-learning approach to predict outbreaks in Chittagong and Cox’s Bazar. Through careful analysis of region-specific survey data, we identified local triggers for dengue spikes. Our data-driven model empowers healthcare officials with predictive ca pabilities to address future outbreaks, safeguarding communities. The study featuresa unique real-time dataset collected from healthcare institutions across Chittagong and Cox’s Bazar, involving collaboration with esteemed public and private institutions. By using diverse data, the dataset aims to unveil hidden insights into dengue outbreaks, guiding accurate predictions and effective prevention measures. The comprehensive pa tient dataset, including diagnoses, medical history, and symptoms, underwent precise model training with a 70:30 split. we have applied various machine learning algorithms, namely - SVM, Decision Tree, XGBoost, Naive Bayes, Random Forest, K-NN, Logistic Regression, and LDA. The Random Forest demonstrated accuracy of 98.92 percent on both training and test data. Performance assessment included a confusion matrixand macro averages for precision, recall, and F1-measure scores.Item “ BANGLA TEXT CLASSIFICATION THROUGH MACHINE LEARNING ALGORITHM”(International Islamic University Chittagong, Department of Computer Science and Engineering, 2023-07) Tarek, Ashraf Uddin; Chowdhury, Satirtha; Hossain, MosharafThe increasing number of social media users and e-commerce platforms have a great impact on people's daily life. People share their emotions and opinion through the social media platform. These emotions and comments now take the important place of analysis because based on these emotions, opinion, business farm make them plan what will produce, consumer decide what will he/she buy etc. Lots of work have been carried out to analysis the sentiment or emotion in English language. Due to the complexity of the Bangla language, few work has been done on it but in recent years several researchers have been carrying out various research based on the Bangla language. In this paper we conduct sentiment analysis (SA) on Bangla language by using the machine leaning model (ML). Most of the work basically divide the sentiment into three categories in this paper we divide the sentiment into four categories namely strong positive, positive, negative, and strongly negative. In this paper we use logistic regression (LR), decision tree (DT), Random Forest (RF), linear support vector machine (LSVM), confusion matrix, and kernel support vector machine (KSVM) algorithm of machine learning (ML). From support vector machine (SVM) we mainly used gaussian kernel radial basis function (RBF). The sentiments are converted to NumPy array to use the sentiment in machine learning. Since the NumPy array is numeric we train our model by these data to get the proper prediction about any given sentiment whether that sentiment positive or negative. The data used are all raw data or primary data collected from the different microblog websites and social media platforms. The total number of raw data 10851 most of them collected from Facebook and YouTube due to their popularity in our country some of the data collected from twitter as well. All the ML model applied for the single word of the sentences first, known as unigram feature analysis best on the single word the logistic regression model and RBF SVM provide the highest accuracy 71.26% and 71.91% respectively. By using two words each model works almost like unigram models, in bigram models LR again shows the highest accuracy, but the accuracy level little bit dropped for RBF SVM. The accuracy for LR and RBF SVM 72.04% and 68.79% respectively. Later we used models for three words defined as trigram feature analysis in that time get highest 70.87% accuracy for LR. Most of the papers basically use SentiWordNet to assess the polarity of the sentiments but in this paper, we use word by word analysis which hardly 6 seen to any paper. This paper will help the business farm as well as the consumers to make their decision and will work as guideline for the new researcher in this topic.Item A grad-CAM and deep learning based image classification for skin diseases(International Islamic University Chittagong, 2022-07) Muntasir, Fahim; Uddin, Mir Arfan; Dey Utsab, Nabajit KumarSkin disease classification and detection are critical and cutting-edge research areas within the field of DeepLearning. With the objective of enhancing the performance of existing di agnostic systems, this study explores the application of deep convolutional neural networks (CNNs), leveraging advanced architectures and visualization techniques. In addressing these challenges, this paper focuses on implementing a customized CNN model while incorporating ResNet-50 and VGG-16 to leverage their combined strengths for enhanced diagnostic accuracy. Additionally, the study employs Grad-CAM for effective visualization of model predictions and decision-making processes.Utilizing the ISIC 2019-2020 dataset, which comprises 9,200 images, our approach encompasses a comprehensive training and testing regimen. The dataset’s diver sity and volume provide a robust foundation for evaluating the efficacy of our proposed models. In parallel, the ISIC Archive dataset further enriches our experimental framework, enabling a nuanced assessment of model performance across varied data scenarios. The custom CNN model attained an accuracy of 87% on the ISIC 2019-2020 dataset, while the CNN+ResNet and CNN+VGG16 configurations achieved accuracies of 95% and 92%, respectively. Regarding the ISIC Archive dataset, the custom CNN model demonstrated a 91.91% accuracy, while the CNN+ResNet and CNN+VGG16 achieved 92% and 94%, respectively.Our findings contribute a novel perspective to the domain of skin disease detection, offering a scalable and efficient frame work that outperforms existing methodologies. The integration of Grad-CAM into our model architecture underscores our commitment to advancing both the accuracy and interpretability of deep learning-based diagnostic systems in dermatology.Item EST: A Semi-supervised ensemble approach for Explainable AI-driven stress detection in social media(Department of Computer Science and Engineering, 2022-12) Rahman, Abdur; Symon, Nurul Karim; Zaman, Akther UzSocial networking websites have become a vast ‘archive’ of human expression, containing hu man emotions from across the globe. Recognizing stress signals within this dynamic landscape is crucial for mental health monitoring, offering opportunities for timely intervention and sup port. This paper introduces a methodology integrating advanced machine learning techniques, notably Ensemble Learning and Self-Training, with Explainable Artificial Intelligence (XAI) to enhance stress detection capabilities. Our research follows a methodical approach, begin ning with foundational steps of data collection and exploratory data analysis, providing insight into sentiments within social media interactions. Prioritizing the interpretability of model pre dictions, our methodology aims to build trust and offer meaningful insights for end-users and mental health professionals. Subsequent phases involve data preprocessing, and refining textual data to extract subtle indicators of stress. The core of our methodology is the fusion of En semble Learning and Self-Training, strategically combining diverse learners to iteratively refine the model using labeled and unlabeled data. With the help of XAI, we seek to deliver clearly and straightforwardly noticeable insights relating to decision-making produced by our stress de tection model. Summing up the presented approach significantly contributes to mental health analytics taking a step forward toward better stress detection by a technical framework with an enhanced accuracy of 91.67%.