Authors - Virgel William Afaga, Patrick Andrew Balang, Dana Wynnette Binwag, Emmanuel Paolo Bromeo, Mark John Bumacod, Carl Allan Calsiman, Juliana April Cendana, Roderick Makil,Dulthe Carlo Munar Jr. Abstract - Benguet Province, Cordillera Administrative Region, Philippines, is highly susceptible to landslides due to its rugged topography, complex geology, frequent typhoon tracks, and extensive mining and road construction. Existing hazard maps rely on static statistical methods and coarse rainfall averages that cannot capture the dynamic triggering conditions of individual storm events. This paper presents a dynamic landslide susceptibility model built on Random Forest (RF) and Extreme Gradient Boosting (XGBoost) trained on thirteen environmental conditioning factors across five domains (topographic: elevation, slope, aspect, distance to streams; geological: rock type, soil type; land cover: LULC, NDVI, NDWI; climatic/hydrological: mean annual rainfall, event rainfall, antecedent rainfall; anthropogenic: distance to roads) derived from high-resolution satellite imagery and event-specific rainfall records. Training used a balanced 16,158-sample dataset (50:50 landslide/non-landslide) from the MGB-CAR geohazard inventory, split 60:20:20 for training, validation, and testing. XGBoost outperformed RF on all metrics: AUC-ROC 0.8903, accuracy 81.78%, precision 81.87%, recall 81.62%, and F1 81.75%; the performance difference was statistically significant (McNemar's test: χ² = 6.22, p < .013). Spatial validation via the Seed Cell Area Index (SCAI) confirmed that High and Very High susceptibility classes captured 69.87% of inventoried landslides within only 36.3% of the provincial area. Expert review by four MGB-CAR geoscientists yielded Likert mean scores above 4.0 for conditioning factor appropriateness, inventory quality, and feature importance plausibility. A fully automated monthly update pipeline was deployed—completing the full cycle from remote-sensing data retrieval to web-map publication in approximately 31 minutes—demonstrating operational feasibility for dynamic hazard monitoring using open-source tools and free satellite data.
Authors - Adibhav Agrawal, Nikunj Parikh Abstract - This article is about how a new configuration of devices has been created for a compact, low-cost, real-time monitoring system for measuring electromagnetic fields on dairy farms. The Electromagnetic Field Monitoring System (EMFMS) is composed of an ESP32 micro-controller, MLX90393 three-axis magnetometer, TP4056 based boosting supply module, and a 0.96-inch OLED screen, which are all encased in a unique 3D printed PETG enclosure. The EMFMS can store and transmit wirelessly time-stamped activity levels of the earth’s magnetic field on three axes through MQTT protocol. The EMFMS was placed into three different areas of an operational dairy barn over 28 days where EMF levels of up to 17 times higher were observed between different areas, and a statistical finding was noted between EMF levels and lower levels of milk production (r = −0.61, p = 0.003) and higher levels of cortisol in serum (r = +0.44, p = 0.03) in Holstein-Friesian dairy cows. The findings of this pilot study demonstrate that this method of continuous measurement of electromagnetic fields for animals using IoT technology could serve as a more feasible and low-cost alternative to existing spot measurements.
Authors - Viraj Bhatt, Rajvi Bhimani, Bhupendra Fataniya, Dhaval Shah Abstract - Cross-border security remains a critical concern for global stability, particularly in jungle or forested terrains where soldiers face significant risks. Military camouflage is engineered to blend in with natural surroundings using advanced concealment techniques that match local textures and color patterns. Consequently, the identification of concealed threats is a challenging task where human observation is prone to error due to poor visibility and fatigue. Traditional surveillance methods often rely on optical sensors which may fail to efficiently detect modern military camouflage. To address this, an automated detection model was developed using the YOLOv8-Nano architecture and deployed on NVIDIA Jetson Nano hardware. The framework was validated using a 5- fold cross-validation strategy to ensure robust and reliable performance. Experimental results yielded a peak mean Average Precision (mAP) at an Intersection over Union (IoU) threshold of 0.5 of 0.955 and an average mAP of 94.8%. The model was further optimized into a TensorRT engine using FP16 quantization, achieving a final footprint of 5.9 MB. These results demonstrate that low-power, portable hardware can effectively perform real-time surveillance as an edge-AI system. This also results in minimizing risks to human lives and directly supporting the core mission of Sustainable Development Goal-16 (SDG-16).
Authors - Kapil Mohan, Ritu Chauhan, Harleen Kaur Abstract - ESG (Environment, Social and Governance) rating in today’s financial world is becoming a good indicator for investors in decision making and risk analysis. There has been stress on E and S in the recent past as Governments and Regulators stress these parameters and benefits to those who are working towards this improvement rating. The rating is a clear indicator of sustainability and promising business and thus is gaining popularity. The analytics firms have combined this indicator and have come up with this calculation using certain scientific and mathematical models from the published data and/or requested data that are provided exclusively to do this calculation for the indicator. These ratings are published annually by analytics firms like Sustain analytics and Bloomberg ESG data service for global but limited firms. This study’s focus is to fit financial data of firms on machine learning models and predict ESG rating with changing market fundamentals and firm’s business value indicators. The result can be com-pared to passed ratings, category averages, deviation and outliers which can benefit venture capitalists and investors to refine their investment strategies. The re-search also captures and compares this output and suggests the approach that best suits this problem by building an architecture that can update the model and can predict live data from the market.
Authors - Anuj Kothawade, Ishan Patra, Pravin Chavare, N. S. Shirude Abstract - The rapid digital transformation of higher education emphasizes the need for robust, data-driven platforms to monitor and enhance student learning. However, many institutions rely on closed, third-party learning management systems that restrict direct access to raw educational data and limit customized analytical capabilities. To address this gap, this paper proposes a scalable educational assessment and learning analytics platform that grants educators complete data sovereignty. Built on a modern stack of TypeScript, React and Tailwind CSS over an owned, directly accessible Firebase backend, the system enables secure, unhindered access for granular data mining. The platform monitors a range of college assessment activities, targeting quizzes and practical coding tests, and uses role-based authentication and custom data-fetching hooks to process student interactions into comprehensive performance metrics. A distinguishing feature is its integrated client-side PDF generation, which instantly produces detailed analytical score reports that serve a dual pedagogical purpose: empowering teachers with actionable insights to adapt instruction, while giving students personalized, self-reflective feedback for continuous improvement. Validated on a controlled pilot, the system achieved 95% overall accuracy, an 85% quiz-evaluation accuracy, a 28% improvement in student engagement, a 40% reduction in report-generation time, and a 92% system-usability score.
Authors - Aleta Fabregas, Nathanael Almazan, Jordan Garcia, Shaina Laman, Paolo Morato, Armin Coronado, Montaigne Molejon, Mariel Leo Violeta Abstract - The Philippines is frequently affected by tropical storms, typhoons, and flooding events that threaten communities located near major river systems. Accurate river level forecasting is essential for improving disaster preparedness and reducing flood-related risks. This study proposes RIVERCAST, a forecasting system that utilizes an Auto-Regressive Transformer with Kernel Principal Component Analysis (Kernel PCA) and Euclidean Kernel to predict Marikina River water levels across the Nangka, Sto. Niño, and Montalban monitoring stations. Meteorological, hydrological, and topographical datasets were collected from PAGASA, MMDA, DPWH, and OpenWeather API records from January 2012 to January 2023. Eighty percent of the collected records were allocated for training while the remaining twenty percent were utilized for testing. The pro-posed model was compared with the Transformer model developed by Xu et al. (2023) using rolling window testing and mean absolute error analysis. Results revealed that the proposed Auto-Regressive Transformer with Kernel PCA and Euclidean Kernel achieved an overall forecasting accuracy of 93.19%, outperforming the Bidirectional Transformer model, which obtained 92.57% accuracy. Findings further indicated that precipitation, rainfall intensity, and temperature significantly influenced forecasting performance, while humidity exhibited the least contribution. The developed model demonstrated reliable twelve-hour river level forecasting capability and may support flood preparedness and early warning initiatives within flood-prone communities along the Marikina River.
Authors - Syafira Aulia Azzahra, Christina Angelica Himawan, Brigitta Vellia, Indra Kusumawardhana Abstract - The hospitality industry is increasingly shaped by smart technology, integrated systems, and subscription-based service models that require consistent and personalized guest experiences. In Indonesia, particularly in the Greater Jakarta area, hotels are adopting Internet of Things-based devices, cloud-based property management systems, and data-driven service platforms to improve guest convenience and strengthen customer retention. This study examines the effects of smart technology devices and integrated systems on customer satisfaction in subscription-based hospitality, with service personalization positioned as a central mechanism in the guest experience. A quantitative cross-sectional survey was conducted with 400 hotel users in the Greater Jakarta area who had experience using smart technology in hotel services. The data were analyzed using Partial Least Squares Structural Equation Modeling. The findings show that smart technology devices and integrated systems positively influence customer satisfaction and service personalization. Service personalization also emerges as the strongest predictor of customer satisfaction, indicating that technology creates greater value when it enables relevant, adaptive, and individualized services. The study contributes to hospitality technology and customer intelligence literature by explaining how digital infrastructure and system integration support personalized subscription-based hotel experiences. Practically, the findings suggest that hotel managers should prioritize investment in interoperable systems, guest data integration, and personalization capabilities to improve satisfaction and sustain long-term customer relationships in technology-enabled hospitality services.
Authors - Aye Mya Mya Win, Ah Nge Htwe Abstract - In recent years, optical flow-based deep learning methods have pro vided evidence of impressive performance in recognizing human behavioral movements from video sequences, revealing high applicability for fall detection functions. This paper analyzes GMFlow-based architectures by experimenting with three different approaches that merge TCN, Attention, and CNN compo nents. These methods are GMFlow-TCN, GMFlow-TCN-Attention, and GMFlow-CNN-TCN-Attention. The experiments were executed on URFD Da taset, Le2i Dataset, and a combined, URFD-Le2i dataset to analyze and evalu ate their performance. According to the experimental results, the method that combines GMFlow-CNN-TCN-Attention achieved better performance than the other proposed models. This model obtained test accuracies of 100% on the URFD dataset, 92% on the Le2i dataset, and 94% on URFD-Le2i dataset. These results point out that the presented method is capable of effectively cap turing both spatial features and temporal features required for fall detection. This approach provides useful insights for developing effective real-time vi sion-based fall detection applications.
