International Journal of Artificial Intelligence & Mathematical Sciences

Data Mining in Healthcare: An Overview of Applications, Techniques, and Challenges

Naseer Ahmed, Rustum Ameer, Afrasiyab Khan, Nooruddin SN, Saba Gull — Wed, 19 Feb 2025 14:08:42 +0000

Artificial intelligence (AI) and the widespread use of electronic health records (EHRs) are driving a data-driven revolution in the healthcare sector mining techniques are extremely useful for deriving significant insights from the large and intricate datasets shaped in healthcare environments. The state-of-the-art uses of data mining in healthcare are examined in this paper, with a focus on AI and machine learning for personalized medicine, risk assessment, disease detection, and therapy optimization. We explore the difficulties that come with data mining for healthcare, including the requirement for interpretable models, data quality, heterogeneity, and privacy data mining and artificial intelligence work together in a way that has the potential to completely transform healthcare delivery, improving patient outcomes, cutting costs, and speeding up medical research. But it's still critical to address ethical issues and make sure data usage is transparent while recognizing the challenges involved, this study emphasizes the revolutionary potential of data mining in healthcare and offers insightful advice for practitioners and scholars in this quickly developing sector.

Analyzing Student's Emotions in the Classroom: A Deep Learning Approach to Facial Expression Recognition

Wed, 19 Feb 2025 14:20:42 +0000

Student success is a crucial metric for educational institutions, as it affects not only the individual student's academic and personal growth but also the institution's reputation, funding, and overall educational outcomes. Student failure is a common and concerning issue in educational institutions, often leading to significant academic and personal consequences. This research explores the potential of deep learning techniques to accurately predict student emotions and prognostication against failure, enabling proactive interventions. Prior studies in this field have mostly depended on conventional machine learning techniques to predict student performance and detect student risk, such as logistic regression and linear support vector machines. However, advanced techniques are frequently needed to capture the underlying patterns and correlations due to the complexity and richness of student behavior and performance. Within the field of artificial intelligence, deep learning has proven to be extremely effective in some applications, such as computer vision, generative AI, and natural language processing. Toward this goal, this research offers an extensive deep learning architecture that utilizes the vast data from the class environment to forecast student emotions more precisely and finely. For evaluating the above facts, the deep learning model CNN-LSTM is used to reveal an effective participation evaluation of students' attention and involvement during classroom sessions. By creating and assessing cutting-edge AI models that improve the precision, flexibility, and real-time capabilities of teaching aids, this research seeks to overcome these constraints. Furthermore, the proposed model demonstrates an accuracy of 91% making it a highly useable real-time application.

Evaluating the Efficiency of Schools in Gilgit-Baltistan: An Exploratory Study

Bulbul Jan, Wajid Ali, Zahoor Hussain, Muhammad Yonus, Faisal Nawaz, Azhar Iqbal — Wed, 19 Feb 2025 00:00:00 +0000

It is known that a well-established education system of a country contributes positively to both the citizens of a nation and its overall economic development and growth. The objective of the study is to identify efficiency and faults in the education sector of Gilgit-Baltistan. For this purpose, this study intent to estimates and evaluates the performance of 1149 government girls and boys’ schools in seven districts of Gilgit-Baltistan using data envelopment analysis (DEA) which helps to highlight ways to reduce inputs or increase outputs necessary for insufficient schools. The utilized data was taken from Directorate of Education Gilgit-Baltistan for the year 2012 to 2013. The results indicate that while the majority of government schools in the Gigit-Baltistan are efficient, a number of them lack in performance due to inadequate use of the available resources. Furthermore, girl’s schools are more efficient as compared to boys' schools in the studied region. In short, this sort of research will be helpful in improving best performance in educational sectors in both government and private sectors.

Artificial Intelligence (AI) Integration in Khmer Chess (Ouk Chaktrang) Game Development

Taing Heangleng, Luy Mithona, Sek Socheat — Wed, 19 Feb 2025 14:50:03 +0000

This paper focuses on developing an AI-based Khmer Chess (Ouk Chaktrang) game using the Python language, without relying on external libraries or frameworks. The project aims to create a Python package that can be used to develop various platforms, providing an accessible and user-friendly game that highlights the unique aspects of Khmer chess, while also serving as a foundation for future AI development in this area. The core of the AI implementation employs the Minimax algorithm combined with alpha-beta pruning to enable the AI to make strategic decisions by evaluating possible moves. Key strengths of this project include a functional game engine with features like move generation and check detection using an efficient reverse-check method. The development uses a 2D array for board representation and implements a piece scoring system. The study also demonstrates its versatility by being implemented in both console and GUI versions. However, key limitations and areas for improvement are identified. The analysis shows a significant increase in computational time as the AI's search depth increases, particularly at higher depths. The AI's move selection is currently random from the list of best moves, leading to unpredictable decisions. Moreover, the AI does not employ advanced techniques such as iterative deepening, selective search, or machine learning due to a lack of training data. In conclusion, this project serves as a foundational benchmark for future development in Khmer Chess AI and aims to promote Cambodian cultural heritage by showcasing the traditional game of Ouk Chaktrang. By addressing the limitations and implementing the proposed enhancements, the project has the potential to become an even more robust and engaging tool for preserving this culturally significant game.

Enhanced Sketch Recognition via Ensemble Matching with Structured Feature Representation

Wed, 19 Feb 2025 15:00:34 +0000

This research explores the use of advanced ensemble techniques including boosting, stacking, hierarchical ensembles, and Bayesian model averaging to improve design recognition performance. Using structural features and combining multiple classifiers, the proposed approach captures complex patterns and hierarchical relationships in design data. Experimental evaluations showed significant improvements against baseline methods, achieving classification accuracy of 92.5%, precision of 91.8%, recall of 90.6%, and F1 score of 91.2%. Boosting and stacking proved to be very effective in capturing complex data features, while hierarchical ensembles effectively handled layer dependencies. The averaging Bayesian model improved reliability by providing a robust uncertainty estimate. Challenges such as computational complexity and dataset balance issues were identified, along with recommendations to improve dataset ordering. These results highlight the potential of combined techniques in advancing contour recognition and provide valuable insights for future research and practical applications in computer vision and pattern recognition.