DSpace-CRIS

Now showing 1 - 3 of 3

Some of the metrics are blocked by your
consent settings
An AI-integrated framework for sustainable conflict mitigation and policy innovation
(2026-09-01)
Tissamana, Apinya
;
Ajayi, Babatunde Oluwaseun
;
Bamisaye, Mayowa Emmanuel
;
Katerenchuk, Wendell
;
Aziz, Tamoor
Conflict in Thailand's peripheral regions remains a persistent challenge, particularly in provinces such as Ubon Ratchathani, where inequalities, cultural marginalization, and resource pressures intersect to shape local tensions. Despite growing attention to conflict analysis, existing approaches often struggle to combine contextual understanding with predictive capability, limiting their usefulness for timely and effective policy intervention. To address this gap, this research introduces a hybrid framework integrating the Decision-Making Trial and Evaluation Laboratory (DEMATEL) method with the Random Forest (RF) algorithm. Using primary data from 400 respondents across urban and rural communities in Ubon Ratchathani, the framework combines stakeholder-informed causal mapping with data-driven prediction. DEMATEL identifies and structures relationships among key conflict dimensions, while RF evaluates and predicts these relationships using empirical data. The findings reveal that cultural and identity issues are the primary drivers of conflict, shaping political and economic marginalization, whereas resource-related conflicts emerge as downstream effects. The RF model demonstrates excellent predictive performance (MSE = 0.0120; RMSE = 0.1097), indicating that these relationships can be reliably translated into predictive insights. This enables policymakers to move beyond reactive responses toward early identification of conflict risks and more targeted interventions. Effective interventions, however, must be coordinated across sectors and tailored to local contexts to achieve lasting impacts. By linking causal understanding with predictive capability, the proposed framework offers a practical tool for conflict monitoring and more inclusive governance aligned with the Sustainable Development Goals (SDGs). While promising, the framework should be tested in other regions to assess its broader applicability.
Some of the metrics are blocked by your
consent settings
Custom Deep Learning Framework for Interpreting Diabetic Retinopathy in Healthcare Diagnostics
(2026-04-01)
Aziz, Tamoor
;
Charoenlarpnopparut, Chalie
;
Mahapakulchai, Srijidtra
;
Ajayi, Babatunde Oluwaseun
;
Bamisaye, Mayowa Emmanuel
Diabetic retinopathy is a prevalent condition and a major public health concern due to its detrimental impact on eyesight. Diabetes is a root cause of its development and damages small blood vessels caused by prolonged high blood sugar levels. The degenerative consequences of diabetic retinopathy are irrevocable if not diagnosed in the early stages of its progression. This ailment triggers the development of retinal lesions, which can be identified for diagnosis and prognosis. However, lesion detection is challenging due to their similarity in intensity profiles to other retinal features, inconsistent sizes, and random locations. This research evaluates a custom deep learning network for classifying retinal images and compares it with the state-of-the-art classifiers. The novel preprocessing method is introduced to reduce the complexity of the diagnostic process and to enhance classification performance by adaptively enhancing images. Despite being a shallow network, the proposed model yields competitive results with an accuracy of 87.66% and an F1-score of 0.78. The evaluation metrics indicate that class imbalance affects the performance of the proposed model despite using the weighted cross-entropy loss. The future contribution will be the inclusion of generative adversarial networks for generating synthetic images to balance the dataset. This research aims to develop a robust computer-aided diagnostic system as a second interpreter for ophthalmologists during the diagnosis and prognosis stages.
Some of the metrics are blocked by your
consent settings
Efficient AI-driven allegation screening: A case study of Thailand’s National Anti-Corruption Commission
(2026-01-01)
Sereewatthanawut, Issara
;
Sriphon, Patipan
;
Khunwipusit, Pattrawut
;
Ajayi, Babatunde Oluwaseun
;
Ilesanmi, Ademola Enitan
Efficient screening of corruption allegations is crucial for promoting accountability and transparency in public administration. However, many institutions still rely on manual processes that are prone to inefficiency and inconsistency. As AI gains traction across sectors, this study develops and evaluates an artificial intelligence (AI)-powered prototype designed to support the preliminary screening of corruption complaints at Thailand’s National Anti-Corruption Commission (NACC). The proposed system integrates Optical Character Recognition (OCR), Natural Language Processing (NLP), and machine learning techniques to automate document handling and improve workflows. A mixed-methods research approach was adopted, combining institutional process analysis with a comprehensive technical performance assessment. The OCR module achieved an F1-score of 81.8%, with precision and recall of 84.2% and 79.6%, respectively. For printed text, the system attained 72% word-level accuracy and 78% at the character level. Additionally, the integrated framework demonstrated a classification accuracy of 57.5% and significantly improved operational efficiency, reducing average complaint processing time by 78.6% compared to traditional manual methods. The findings highlight AI’s transformative potential in enhancing anti-corruption efforts through increased speed, accuracy, and consistency. They underscore the importance of responsible and context-sensitive AI adoption in public sector governance. This study contributes to the growing discourse on digital governance by providing empirical evidence and practical insights for policymakers and practitioners aiming to implement scalable, transparent, and ethically grounded AI solutions within institutional accountability frameworks.

KMITL

Browse

Filters

Settings

Sort By

Results per page

Search Results