Rigs
James Henry
Hadayat Son
Exam Analyzer
Developed an Exam Analyzer for educational institutions to identify poorly designed questions causing a high failure rate. Analyzed student responses and key files to prepare results, focusing on the correlation between top students' performance and the average performance of all students on each question. This project aimed to enhance the quality of assessments by pinpointing and addressing problematic questions.
Opinion with Opinion: Segmentation Approach for Urdu Sentiment Analysis
Developed a segmentation approach for Urdu sentiment analysis, identifying opinion segments within text to improve sentiment classification accuracy.
Predicting Mental Illness from Twitter Activity
Utilized activity theory and context ontology to predict mental illness from Twitter activity, providing insights into behavioral patterns associated with mental health issues.
Social Media News Classification in Healthcare Communication
Classified healthcare-related news on social media using machine learning techniques to improve the dissemination of reliable health information.
Design Quality Metrics for Autonomic Computing Systems
Developed design quality metrics to evaluate the suitability and cost-effectiveness of self-capabilities in autonomic computing systems.
Upgrading Legacy Cameras for Threat Monitoring
Engineered a cost-effective solution for upgrading legacy cameras to monitor and track live threats using software and hardware integration.
Self-Risk Assessment for Cervical Cancer
Developed a machine learning based self-risk assessment technique for cervical cancer, providing personalized risk predictions.
Gas Consumption Analysis of Ethereum Blockchain Transactions
Conducted an analysis of gas consumption in Ethereum blockchain transactions to optimize resource usage.
Multiagent Survival in Social Dilemmas
Improved the survival time of multiagents in social dilemmas using a neurotransmitter-based Deep Q-Learning model of emotions.
Quantum Variational Circuit for Resource Management
Introduced a quantum variational circuit for efficient management of common pool resources.
Seed Clustering and Visualization using PCA
This project assigned various seed types into distinct clusters and visualized them in two dimensions. Using features like area, perimeter, compactness, kernel length, kernel width, and asymmetry coefficient, we determined the optimal number of clusters with the elbow method and WCSS. We applied K-means clustering and used PCA to reduce data dimensions, preserving variance and structure for clear visualization. Technologies used included Python, Pandas, NumPy, K-means, PCA, Matplotlib, and Scikit-learn.
Penguin Classification Using Logistic Regression
This project developed a multi-class classification model to classify penguin species using a logistic regression classifier. We cleaned the dataset, removed null values, and handled missing data to ensure quality input. The model was trained with optimized hyperparameters and evaluated using metrics like ROC, AUC, precision, recall, accuracy, and F1 score. Technologies used included Python, Logistic Regression, Scikit-learn, Pandas, NumPy, and Matplotlib.
Diabetic Data Analysis and Classification
This project identified diabetic patients using logistic regression, decision trees, and ensemble learning. We selected features with box plots, trained models with various techniques, and evaluated performance using precision, recall, and F1 score. Technologies used included Python, Logistic Regression, Decision Trees, Ensemble Learning, Scikit-learn, Pandas, NumPy, and Matplotlib.
Daily Bike Share Rental Prediction
This project focused on predicting bicycle rentals for a bike-sharing company based on weather conditions. Multiple regression models were developed and trained, including linear regression, decision tree regressor, ensemble learning regressor, and gradient boosting regressor. These models were evaluated and compared using the coefficient of determination (R²) and sum of squared errors (SSE). The technologies used included Python, Scikit-learn, Pandas, NumPy, and Matplotlib.
Movie Rating Prediction
This project focused on predicting bicycle rentals for a bike-sharing company based on weather conditions. Multiple regression models were developed and trained, including linear regression, decision tree regressor, ensemble learning regressor, and gradient boosting regressor. These models were evaluated and compared using the coefficient of determination (R²) and sum of squared errors (SSE). The technologies used included Python, Scikit-learn, Pandas, NumPy, and Matplotlib.