M.Phil / MS

Now showing 1 - 3 of 3

Machine Learning Approaches in Medical Diagnosis
(Library Information Services, COMSATS University Islamabad, Lahore Campus, 2025) Muhammad Bilal; CIIT/SP24-RMT-001/LHR; Dr. Muhammad Yousaf Bhatti; LHR TP 10065
Brain tumors are among the most life-threatening neurological conditions, characterized by abnormal and uncontrolled cell growth within brain tissues. Early identification of tumor type and stage is crucial for effective intervention, yet conventional diagnostic methods often struggle to detect tumors in their early or subtle stages. This study proposes advanced deep-learning models for multi-class classification of brain tumors and machine-learning approaches for predicting chemical properties of anti-cancer drugs to support new drugs development. A hybrid deep-learning model was developed for classifying three tumor categories glioma, Meningioma and pituitary tumors using MRI scans. the model was trained on the brain tumor MRI dataset and integrates ResNet50 and Efficient Net50 as its base ar-chitectures, combined through a novel triangular dense-layer fusion strategy to optimize multi-stage feature extraction. The proposed model achieved a test accuracy of 97%, recall of 97%, and an AUC score of 99%, demonstrating its effectiveness for early and accurate brain-tumor classification. In second phase of this research, machine-learning algorithms and topological indices were used to analyze molecular structure of thirteen anti-cancer drugs. Topological in-dices were computed using a python program, and true physicochemical properties were retrieved from the ChemSpider database via an automated script. QSPR and SHAP anal-ysis were performed to identify indices most predictive of each physicochemical property. Machine-learning models were then trained on these features to develop generalized pre-dictive models for key durg properties, providing a foundation for computationally assisted drug discovery. Overall, this thesis demonstrates the combined potential of deep learning for early tu-mor detection and machine-learning for anti-cancer drug analysis, ultimately contributing to improved diagnostic accuracy and future therapeutic development
Connection Numbers on Carbon Nanocone Line Graph and Various Chemical Lin
(Library Information Services COMSATS University Lahore Campus, 2023-03-13) Muhammad Bilal; Dr.Imran Zulfiqar Cheem; SP22-RMT-024; LHR TP 8730
In this study, we delve into the mathematical exploration of the Connection Number in carbon nanocone line graphs and diverse chemical line graphs within the realm of Graph Theory. The theoretical framework is developed based on the principles of connectivity and network structures. These structures are represented as graphs, where atoms con stitute nodes and chemical bonds form edges. The novel metric, Connection Number, is introduced to quantify connectivity within these unique molecular configurations. Em ploying graph invariants and specialized connectivity indices, the study characterizes and compares the structural features of carbon nanocones and various chemical line graphs. Advanced graph algorithms are utilized to unveil the intricate relationship be tween Connection Number and structural stability, offering insights into the impact of connectivity on molecular properties and reactivity. The analysis extends to identifying recurring structural motifs and patterns, contributing to a systematic classification of connectivity profiles. Beyond theoretical considerations, the study explores practical applications in nanotechnology, materials science, and chemical engineering. This con cise investigation harnesses the power of graph theory to deepen our understanding of molecular connectivity in carbon nanocones and diverse chemical line graphs, paving the way for innovative advancements in nanoscience and materials engineering
On Cluster Analysis of Some Portfolio Optimization
(Library Information Services, COMSATS University Islamabad, Lahore Campus, 2023) Muhammad Bilal; CIIT/SP22-RMT-009/LHR; Dr. Sarfraz Ahmad; LHR TP 8716
The classic mean-variance portfolio optimization approach is criticized in large part for its propensity to overstate estimate error. An estimated inaccuracy of a few percent can skew the entire package. The Black-Litterman technique (Bayesian method) and the resampling methodaretwocommonapproachestosolvingthisproblem. Amorerecentapproachtothe issue’s solution is the clustering method. By clustering, we initially combine the stocks that have a strong correlation and handle the group as a single stock. Following the grouping of the stocks, we will have a few stock clusters. For these clusters, we do the standard mean-variance portfolio optimization. By using the clustering approach, the influence of estimating error may be minimized and the portfolio’s stability can be increased. In this project, we’ll examine how it functions and run experiments to see if clustering techniques enhance the portfolio’s performance and stabilities

M.Phil / MS

Browse

Filters

Settings

Sort By

Results per page

Search Results