UPLB ICS Peak One

Financial literacy remains a pressing issue among university students in the Philippines, who often face financial challenges due to limited resources and rising living costs. Previ- ous approaches to student budgeting have often relied on static tools or generic financial advice, which fail to account for indi- vidual financial behavior and changing economic circumstances. This study introduces ForeSight, a financial management system that uses the Random Forest algorithm to provide personalized budget forecasts based on sociodemographic data and historical spending patterns. Previous tools lacked personalization and adaptability; ForeSight addresses this by integrating predictive analytics with features such as transaction tracking, goal setting, and data visualization. Usability evaluation using the System Us- ability Scale (SUS) and user feedback revealed high satisfaction, particularly in ease of use and relevance of recommendations. Findings demonstrate that machine learning can significantly enhance financial decision-making, offering a scalable approach to improving financial literacy among students.

The University of the Philippines Open University offer online courses that is available for people around the world. This means that the instructors needs to deal with thousands of students for their respective courses. Addressing each students’ questions is impossible. OPENG is a course level chatbot developed to assist the instructors in this regard. Students are given the platform to ask course related questions to the chatbot and receive instant feedback. The results of the testing showed that the chatbot offers a great user experience, being easy to use and understand. Furthermore, testing showed that the chatbot responded with the correct answers, given that the questions were inside the scope of the course.

Phishing attacks on social media platforms have become increasingly sophisticated, posing significant threats to users’ personal information and online security. To address this, a real-time phishing detection system was designed specifically for URLs shared on Facebook. The system employs a machine learning approach using a Random Forest classifier trained on a dataset of 396,730 URLs, achieving 85.06% accuracy after hyperparameter tuning. Feature importance analysis revealed that URL structural characteristics, particularly length-based features, are the most reliable indicators for phishing detection. Performance testing demonstrated efficient response times with a median of 256.92 ms and reliable scalability for concurrent users. A System Usability Scale evaluation with 22 participants yielded an exceptional score of 92.73, confirming the exten- sion’s effectiveness and user-friendliness. Despite the reliable performance, opportunities for improvement exist, particularly in prediction accuracy and response time optimization. DeBaiter is publicly available through the Chrome Web Store and represents a significant advancement in real-time phishing protection for social media users.

The advancement of technology presents significant opportunities for improving the healthcare sector, particularly in the area of patient information management. This study focuses on Meycauayan Eye Clinic, a healthcare facility that previously relied solely on Google Sheets to store patient data which hindered effective tracking of patient history. Moreover, manually notifying patients of upcoming appointments posed a considerable burden on the staff. To address these issues, this study developed a web application named WebSight to digitize the clinic’s database and integrate an SMS notification feature, thereby reducing the clinic staff’s workload. The application was built using the MERN stack (MongoDB, Express.js, React, and Node.js), alongside TextBee.dev for the SMS service. Access to the system is restricted to three user roles: Admin, Staff, and Doctor. For security and privacy, only assigned usernames and passwords are allowed, Gmail or other personal email logins are not permitted. Each user role is provided with dedicated features and functionalities. The usability of the system was evaluated using the System Usability Scale (SUS), yielding an average score of 87.5user satisfaction and ease of use.

Movies have long been a prominent form of en- tertainment, greatly influencing humanity’s perspectives, beliefs, and values. Philippines, with its rich cultural diversity, embraces movies as it reflects the nation’s social, political, and economic landscapes, depicting traditions and contemporary social issues relevant to Filipinos. With an ever-growing catalog of Filipino films, viewers often face the challenge of finding content that resonates with their individual tastes and preferences. This project introduces a Filipino movie recommendation system that delivers personalized suggestions using three weighted recom- mendation techniques: Content-Based Filtering which uses movie information, User-Based Collaborative Filtering which uses user ratings and similarity, and Knowledge-Based Filtering which uses movie ratings and popularity. By utilizing these techniques, the platform offers users a dedicated space to discover Filipino movies in a way that feels both intuitive and personalized. We also evaluated the effectiveness of each recommendation approach, comparing their ability to deliver relevant suggestions. Results show that Content-Based Filtering is most effective in determining movies that users have watched with 289 movies and Knowledge-Based filtering has the highest positive rating score with 87.26%.