eScholar-UMT
eScholar is the institutional repository for research conducted at UMT and maintains a large collection of theses, dissertations and projects produced by UMT graduates as part of their respective degree programs. It includes (but not limited to):
- PhD/MS Theses
- Graduate Program Research Projects
- Undergraduate Program Reports and Final Year Projects
- Full-text articles/research work of faculty and students
Join eScholar and sign-up to get the access of above mentuobed digital resources at your desktop.
Communities in DSpace
Select a community to browse its collections.
Recent Submissions
AI Based Cyber Attacks Detection Model for IoT Networks
(UMT, Lahore, 2025) Hina Jabbar
With the Passage of time adoption of IoT continues rise the threat of cyber-attacks is also growing. It demanding the effective and accurate mechanism for detection. Traditional cyber attack detection mechanisms often suffer from the imbalanced attacks classification, underutilization of the datasets and high false negative rates especially for the minority attack categories. These limitations decrease the ability of models to detect less frequent but most critical types of attacks, compromising cybersecurity. This study addresses these challenges by familiarizes an optimized model based on the Gradient Boosting for cyber attacks detection. The model is design to enhance the minority attacks classes while maintaining the overall accuracy. To attain this, we employ CICIoT2023 dataset, utilizing its large-scale structure to ensure the comprehensive model training and robust generalization of it. The evaluation of large-scale dataset provides better generalization of model that improved the representation of the real-world pattern of attack and it help to reduce bias. It allowing the model to make classification more accurately.so, various preprocessing techniques including data resampling and dimensionality reduction applied to model learning and address the class imbalance challenge. However, the resampling methods help balance the classes of dataset but lead to overfitting, generate artificial pattern and decrease ability of the model to generalize to unseen the attacks. We introduce multiple variants of the Gradient Boosting model through tunning of hyperparameter. One optimized variant GB_10D4 demonstrating best performance both for binary and multiclassification.
PhageVir: A Machine and Deep Learning Approach for Effective Prediction of Phage Virion Proteins
(UMT, Lahore, 2024) HUSSNAIN ARSHAD
This study uses machine and deep learning techniques to create an effective prediction model for PVPs. PVPs are key phage components that aid in the attachment and penetration of the phage to the host cell, allowing viral reproduction. PVP prediction accuracy can provide crucial insights into the molecular mechanisms of phage infection and may contribute to developing novel antibacterial treatments. Various state-of-the-art machine and deep learning approaches were used: including LGBM, RF, CNN1D, LSTM, RNN, GRU, and ANN to identify PVPs. Existing literature fervently supports the use of these algorithms based on recent bioinformatics studies, as they have proved helpful in processing sequential data. Precision metrics, like accuracy, sensitivity, specificity, and MCC, were calculated to assess the models' performance. Each model underwent independent testing, self-consistency, and cross-validation to achieve accurate findings. The results of this study demonstrate that machine and deep learning techniques accurately predict PVPs. The highest 10-fold cross-validation accuracy, sensitivity, specificity, and MCC score were achieved by CNN attaining 0.833 accuracy, 0.832 sensitivity, 0.834specificity, and an MCC score of 0.665. The ROC curve also revealed CNN performed well, achieving an AUC score of 0.927. Based on rigorous experimental evidence, it is inferred that the work proposes effective machine and deep learning techniques to classify PVPs accurately. The web server has been deployed at https://hussnain-arshad-phage virion.streamlit.app/
Forecasting Particular Matter (PM 2.5) Air Pollution Trends using Time Series Analysis based on Hybrid Arima-Prophet Models
(UMT, Lahore, 2024) SAMIYA SHARIF
Air quality in Pakistan is a major public health concern, with Particular Matter (PM) 2.5 m(µg/m³), a particularly harmful fine particulate matter, posing significant health risks. However, effective monitoring of air quality faces significant challenges. This research delves into these challenges and explores potential solutions for improved forecasting of Particular Matter PM2.5 (µg/m³) levels. One major obstacle is technical issues at air quality monitoring stations. Air Pollution data using time series analysis techniques. A time series analysis conducted the air pollution dataset of four cities i.e. Lahore, Islamabad, Karachi and Peshawar to forecast future trends from 2020 to 2024. Four models were used, including ARIMA, Prophet, LSTM, and the proposed Hybrid Arima Prophet model. Proposed Hybrid Arima-Prophet Model, a combination of ARIMA and Prophet, was found to be the most accurate in terms of prediction accuracy, Peshawar achieved the lowest RMSE (0.0043) and MSE (0.0506), indicating the most accurate forecasts. Islamabad, Lahore, and Karachi followed with increasing RMSE (0.0145,0.0065, and 0.0480) and MSE (0.0972, 0.0646, and 0.1764) values, respectively. The LSTM model was not suitable for this type of data with a high RMSE and MSE. This study highlights the potential of combining models in time series analysis and the importance of considering multiple models in forecasting future trends. This Work provide new insights into the temporal patterns of air pollution data and can contribute to the development of more effective strategies. However, it is important to note that these results are specific to the data used. Further research is needed to explore the impact of external factors like weather and traffic on air pollution levels.
AI – Driven MCQS Generation Using LLM
(UMT, Lahore, 2024) Muhammad Qamar Iqbal
Multiple choice questions (MCQs) in educational assessment, are essential because they provide a systematic way to gauge scholars' comprehension and knowledge of a wide range of subjects. This paper investigates the novel use of Artificial Intelligence( AI)- driven Large Language Models (LLMs) to automate the generation of multiple- choice questions (MCQs) for use in educational evaluations. A novel approach to expedite the creation of multiple choice questions (MCQs) while maintaining their validity and efficiency is presented, utilizing state of the art NLP techniques and merging them with LLMs. The AI- driven system is methodologically trained and validated by precisely utilizing a wide range of educational resources and sources. The system's ability to produce multiple- choice questions (MCQs) that are both accurate and contextually applicable is demonstrated through expansive trial, encompassing a wide range of subjects and difficulty situations. This framework is unique because of its personalized approach, which enables MCQs to be tailored to each student's specific requirements and learning preferences, fostering adaptable learning environments. The ramifications of AI- driven MCQ product on assessment procedures are also covered in the discussion, with a focus on how crucial it's to preserve inclusivity, availability, and fairness. In general, this study provides insightful information about the relationship between artificial intelligence (AI) and educational evaluation, opening doors for inventions in pedagogy and personalized learning.
Dilemmas of the Pak-India Indus Water Settlement
(UMT.Lahore, 2017-06-05) NAZIRULLAH