Computer Sciences

Permanent URI for this collectionhttps://repository.aab-edu.net/handle/123456789/15

Browse

Search Results

Now showing 1 - 10 of 43
  • Item
    Dizajnimi dhe implementimi i rrjetit Wireless në kompaninë Asseco
    (2026-10-19) Kele Kiseri, Fidan
    Internet users are constantly increasing in line with growth of population and access to the Internet in an easier way. The easiest way is to use the Internet not only from your desktop or personal computer that is connected to the Internet with network cable, but nowadays you can roam with laptops, mobile phones, iPad, etc. The Internet may be accessible through wireless means of establishing modems or providers of a signal not only through the internet cable network. So, the idea to do this theme of diploma born from the importance of accessibility and popularity of wireless for companies but also for Internet users in homes. The contemporary human society development in the field have increased need for identification of problems according to priorities and determine ways of analysis and management of working in this field. This project includes research, design and implementation of wireless systems for the known company Asseco. The research in this project comprises the researches in hardware and software equipment as well as designing and implementing successful. Requirements to be in step with the competition, they have made the changes to be persistent and have their dominant effects for this type of communication that is always stretch and in the above expansion.
  • Item
    Analiza dhe optimizimi i performancës së rrjetave LAN duke përdorur Cisco Packet Tracer
    (AAB College, 2026-10-19) Shala, Kastriot
    This paper deals with the analysis and optimization of network performance (LAN) and its implementation in Cisco Packet Tracer simulator. The main goal is to find the causes that affect network performance such as: latency, throughput, packet loss and diversity acceptance and the application of network designs to improve efficiency, stability and security. In the tedious theoretical work that is necessary, for the study of the practice of traffic management, the basic standards and protocols (e.g. Ethernet, VLAN, STP) and the exercise with privilege management tools, the basic security measures ACL, Port Security and DHCP Snooping were taken into account. Then, in Cisco Packet Tracer, a virtual environment was created where the LAN version was designed and the addition of VLAN, Traffic Shaping with QoS, Loop Prevention with STP and Ethernet Channel Aggregation (LACP) was implemented. There were implementations and several mechanisms for preventing authorization and for regulating IP conflict. The simulation shows that the application of VLANs and the good configuration of DHCP reduces IP address conflicts and broadcast traffic. The activation of STP for both Ethernet and EtherChannel increases redundancy and reduces latency costs in the environment between switches, while QoS provides stable levels of service for critical services such as VoIP telephony and video conferencing. Ping tests in the core network after optimizations resulted in 0% packet loss in total and latency below 1ms. This paper simplifies that simulations with Cisco Packet Tracer as attacks before planning the behavior of a network in real conditions, are still an idea for the speed of the optimal construction without using gages and the discovery of tested routes and stationary that is suitable for infrastructures that allow to stay fast and fast with security constructions
  • Item
    Testimi i Përdorshmërisë së Sistemeve të Menaxhimit të Mësimit. Rast Studimi: Portali “e-student” në Kolegjin AAB
    (AAB College, 2026-10-19) Adili, Ashkën
    This study evaluated the usability of AAB College’s “e-student” platform, focusing on ease of use, efficiency, satisfaction, communication, and mobile access. A mixed-methods design was applied: a quantitative survey with 35 students, an open-ended qualitative questionnaire with 13 users, a self-reported practical task test with 10 participants, and one semi-structured institutional interview. The data were analysed through descriptive statistics, thematic analysis, and methodological triangulation. The results indicated a generally positive assessment: the mean of the 24 items evaluating the platform’s current state was M=4.15/5, while the overall rating was M=7.84/10 based on 35 valid responses. All seven practical tasks were reported as successfully completed by all 10 participants. Nevertheless, recurring weaknesses were identified in notifications, communication, mobile access, and selected navigation flows. The findings provided an evidence-based foundation for improving the platform and an applicable model for evaluating comparable systems in higher education.
  • Item
    Analiza dhe Segmentimi i Klientëve në Biznes duke përdorur Teknikat e Data Mining
    (AAB College, 2026-10-19) Osmanovic, Bljeard
    This thesis focuses on customer analysis and segmentation using Data Mining techniques in the context of business and sales. The study is based on the analysis of transactional sales data in order to identify customer behaviour patterns and to classify customers according to the recency of their last purchase, their purchasing frequency and their monetary value. The analysis was performed on the publicly available "Online Retail" dataset from the UCI Machine Learning Repository, which contains 541,909 transaction records collected between December 2010 and December 2011. After cleaning, 392,692 records belonging to 4,338 unique customers were retained. The RFM model (Recency, Frequency, Monetary) was combined with the K-Means clustering algorithm, which segmented customers into four groups with distinct characteristics. Agglomerative Hierarchical Clustering with Ward linkage was applied to the same data for comparison. Because the Frequency and Monetary variables are heavily right-skewed (skewness of 12.07 and 19.34 respectively), a logarithmic transformation was applied prior to standardisation. The number of clusters was determined using the Elbow Method and evaluated with the Silhouette Score, which reached 0.3368 for the four-cluster solution. A control analysis without the logarithmic transformation produced a nominally higher score (0.6164) but isolated only 13 extreme customers while leaving 70.5% of the customer base undifferentiated, demonstrating that internal validity indices alone are insufficient for model selection on skewed data. The central finding is a pronounced concentration of value: 699 customers, or 16.1% of the base, generate 64.5% of total revenue of £8,887,208.89, whereas the 1,635 inactive customers who make up 37.7% of the base contribute only 6.3%. The four segments differ systematically across all three RFM dimensions and each can be associated with a distinct marketing action: retention for loyal customers, development for customers with growth potential, activation for low-frequency buyers and re-engagement for customers at risk of churn. K-Means outperformed Hierarchical Clustering on both silhouette (0.3368 vs 0.2974) and runtime, while the substantial agreement between the two partitions (ARI = 0.52) supports the robustness of the structure found. The study demonstrates that relatively simple Data Mining techniques, applied to transactional data that any sales system already records, produce actionable knowledge for marketing and customer management decisions without requiring complex infrastructure.
  • Item
    Sfidat e përdorimit të inteligjencës artificiale në arsimin e mesëm në Kosovë
    (AAB College, 2026-08-21) Buzhala, Rrezarta
    This thesis examines the challenges and perceptions related to the use of artificial intelligence in secondary education in Kosovo. The aim of the study is to analyze the level of knowledge, practical use, and attitudes of students and teachers toward artificial intelligence in the learning process. The research is conducted using a quantitative approach through two structured questionnaires addressed to secondary school students and teachers. The findings indicate that artificial intelligence is widely used, particularly through tools such as ChatGPT, and is mainly perceived as a supportive tool that facilitates learning, lesson planning, and the development of digital skills. However, the lack of professional training and technological infrastructure represents significant challenges for effective integration. The study emphasizes the need for a planned institutional approach and responsible use of artificial intelligence in education.
  • Item
    Analizimi dhe Vizualizimi i të Dhënave me PowerBI në Institucionet e Arsimit në Kosovë
    (AAB College, 2026-08-21) Berisha, Misret
    In the era of digital transformation, data analysis and visualization have assumed a pivotal role in supporting decision-making across various institutions, including the education sector. This paper addresses the importance of data analysis and visualization using the Power BI platform within educational institutions in Kosovo, focusing on leveraging data to enhance institutional management and performance. The paper examines fundamental concepts regarding data analysis, Business Intelligence, data quality, and digital transformation in education. It also outlines the key functionalities of Power BI, the types of visualizations available, and the benefits the platform offers for analyzing and interpreting educational data. The empirical component of the study relies on an analysis of educational data from Kosovo, sourced from the Kosovo Agency of Statistics (KAS) and the United Nations Development Programme (UNDP). Using Power BI, interactive reports and dashboards were created to present data on student demographics—such as gender, ethnicity, and municipality—as well as levels of study, enrollment figures, graduation rates, and dropout statistics. The results demonstrate that the use of data analysis and visualization enables the identification of trends, challenges, and opportunities for improving the education system. Furthermore, Power BI offers an effective approach to transforming data into useful information that supports strategic decision-making and evidence-based management. The paper concludes that integrating modern Business Intelligence tools into educational institutions can significantly contribute to enhancing efficiency, transparency, and the quality of education in Kosovo.
  • Item
    Sfidat e Përdorimit të Teknologjisë së Informacionit në Përkrahjen e Fëmijëve me Aftësi të Kufizuara në Republikën e Kosovës
    (AAB College, 2026-08-21) Mlinaku, Hava
    Information and Communication Technology (ICT) plays an important role in promoting inclusive education and supporting children with disabilities by providing greater opportunities for access to information, communication, and active participation in the learning process.The aim of this study is to identify and analyze the main challenges related to the use of Information and Communication Technology in supporting children with disabilities in Kosovo, as well as to assess the level of technology integration in educational institutions. The findings of the study show that educational institutions have made progress in equipping schools with technology and using it in the teaching process; however, challenges still exist, including the lack of specialized equipment, the need for continuous teacher training, financial support, and the provision of assistants for children with special needs. This study contributes to the enrichment of the literature on the use of ICT in education and offers practical recommendations for educational institutions and policymakers, with the aim of improving access to and use of technology for children with disabilities in Kosovo.
  • Item
    Zhvillimi i një Sistemi për Detektimin e Objekteve Dominuese në Imazhe duke përdorur Deep Learning (Salient Object Detection System)
    (AAB College, 2026-07-22) Mustafa, Shkurte
    This paper addresses the development of a system for detecting dominant objects in images using deep learning and convolutional architectures implemented with PyTorch. Two models were constructed and compared: a baseline CNN model and an enhanced model incorporating Batch Normalization, Dropout, skip connections (based on the U-Net architecture), and advanced data augmentation techniques. The models were trained and tested on the ECSSD and DUTS datasets and evaluated using IoU, Dice, Precision, Recall, F1-Score, and MAE metrics. The enhanced model outperformed the baseline model on both datasets: on ECSSD, the IoU increased from 0.464 to 0.532, while on DUTS it rose from 0.685 to 0.766, alongside improvements in Precision, F1-Score, and Accuracy. The results indicate that the enhanced model produces clearer object boundaries and reduces false positives, although the segmentation of objects with complex shapes remains the primary challenge.
  • Item
    Implementimi i pagesave digjitale në Sektorin Shëndetësor Prezantimi/Demonstrimi i Ueb Softuerit
    (AAB College, 2026-07-22) Dida, Shefqet
    This study addresses the development and evaluation of an integrated system for patient management and the implementation of digital payments in the healthcare sector. The main goal of the paper is to analyze how such a system can improve operational efficiency, reduce administrative time and costs, and increase transparency and security in the management of data and financial processes. The study is based on an applied and experimental approach, where a functional prototype of the system has been developed and tested under controlled conditions. The methodology includes system performance analysis, security assessment, and comparison with traditional management and payment methods. Data were collected through direct testing of functionalities, response time analysis, and evaluation of the system architecture structure. The results show that the developed system significantly improves the efficiency of administrative processes by enabling automation of patient management and billing. The system demonstrates satisfactory performance under normal use conditions and offers higher transparency compared to traditional methods. However, several limitations have been identified, including the lack of real integration with financial systems and the need for improvements in terms of security and scalability. In conclusion, the system represents a promising solution for the digitalization of processes in the health sector, providing a solid basis for further development and practical implementation. The study recommends full integration with payment platforms, improvement of security mechanisms to increase the reliability and applicability of the system.
  • Item
    Zbulimi i sulmeve phishing dhe inxhinierisë sociale dhe vlerësimi i efektivitetit të sigurisë kibernetike në sektorin bankar: Evidencë nga Kosova
    (AAB College, 2026-07-22) Iliyaz, Ertan
    In today’s digital age, social engineering has become significantly more complex, with increasingly sophisticated attacks. This development provides criminals and organized groups with greater opportunities to leverage artificial intelligence to cause even more extensive damage. For example, the use of the Albanian language in phishing attacks is currently observed in Kosovo — a tactic that enables attackers to more easily gain the trust of clients. A vast illegal economy is evolving in the cyber world. This field has grown so rapidly that even young people are becoming increasingly involved in cybercrime. These groups often target elderly individuals through various social engineering methods, with the aim of stealing financial assets from their bank accounts. Consequently, it is essential for both banking institutions and clients to continuously monitor technological developments to ensure a higher level of security. This thesis examines the nature and forms of social engineering and phishing attacks, with a particular focus on the banking sector in Kosovo. The primary objective is to assess the current level of awareness among bank employees and clients regarding these threats, as well as to evaluate the effectiveness of existing protective measures. Through a mixed-methods approach that includes quantitative data from surveys with users of banking services and the analysis of security reports, the study identifies the main weaknesses in the cybersecurity infrastructure. The findings reveal that, despite investments in security technologies, the human factor remains the weakest link in the chain of defense. The use of sophisticated techniques such as spear-phishing and smishing has significantly increased the success rate of attacks. This thesis emphasizes the urgent need for a comprehensive approach that combines advanced technology with continuous education and awareness programs for clients. The final recommendations provide a strategic framework for financial institutions in Kosovo to strengthen their resilience against cyber threats and to better protect the financial assets of citizens. In the digital era of 2026, the banking sector is characterized by two parallel realities: on one hand, rapid digitalization and the expansion of online services; on the other hand, the exponential growth of cyber threats. According to the ENISA Threat Landscape 2025 report, phishing constitutes approximately 60% of the initial attack vectors in the financial sector across Europe. In the Western Balkans, and particularly in Kosovo, this phenomenon is even more alarming due to the use of the Albanian language, familiarity with the local context, and the relatively low level of cybersecurity awareness. Social engineering attacks exploit the greatest vulnerability of any system the human element. Even the most advanced technical systems can be compromised if an employee or client clicks on a fraudulent link. This study aims to shed light on the Kosovo reality by combining the analysis of internal bank data, interviews with professionals, and a review of international literature. The importance of this topic is high because the banking sector is considered critical infrastructure. Any successful incident can result in direct financial losses, damage to customer trust, and regulatory sanctions.