https://repository.umindanao.edu.ph/handle/123456789/2292 2026-05-20T09:46:13Z 2026-05-20T09:46:13Z Jimenez, Cyril Jefferson Calape Magallanes, Jose Fernando Bactol Tabay, Carl James Hermoso Genobiagon, Cresencio Pombo Jr https://repository.umindanao.edu.ph/handle/123456789/2326 2026-05-07T19:00:31Z 2024-01-01T00:00:00Z

Organic phase change materials as thermal energy storage for automated solar seawater desalination system Jimenez, Cyril Jefferson Calape; Magallanes, Jose Fernando Bactol; Tabay, Carl James Hermoso; Genobiagon, Cresencio Pombo Jr A prototype of an automated solar seawater desalination system equipped with thermal energy storage was designed and developed to analyze experimentally the effect of adding organic phase change materials (palm wax and beeswax) to the desalination system’s efficiency. The desalination machine was set up with two configurations: with organic and without organic PCMs. While the setup without a PCM can produce an average of 30.8 mL per day and convert 2.056% of the total volume of input water to an output peak of 48.5 mL using solar energy, the setup with a PCM can produce an average daily water production of 68.9 mL and convert 4.6% of the input water volume with the highest record output of 104.5 mL compared to configurations without PCM, the organic PCM system with a 1:1.2 mixture of beeswax and palm wax can produce an average of 123.7%. A prototype solar seawater desalination system with thermal energy storage was developed using organic phase change materials (palm wax and beeswax). The PCM-enhanced setup significantly boosted efficiency, producing up to 104.5 mL daily and achieving a 123.7% increase compared to the non-PCM configuration.

2024-01-01T00:00:00Z Fajardo, Reban Cliff A. Yara, Fe B. Adeña, Randy F. Hernandez, Michael Kevin L. Arroyo, Jan Carlo T. https://repository.umindanao.edu.ph/handle/123456789/2324 2026-05-07T19:00:43Z 2024-01-01T00:00:00Z

A data-driven approach in predicting scholarship grants of a local government unit in the Philippines using machine learning Fajardo, Reban Cliff A.; Yara, Fe B.; Adeña, Randy F.; Hernandez, Michael Kevin L.; Arroyo, Jan Carlo T. Inefficient, tedious, and outdated processes in resource allocation are some of the common hurdles educational institutions and agencies face in managing scholarship grants and selecting potential grantees. In response to the challenge, this study developed a predictive model utilizing a range of machine learning algorithms; by leveraging algorithms like Naïve Bayes, Random Forest, Logistic Regression, Support Vector Machine, and Multilayer Perceptron, the study aimed to enhance the selection process for scholarship programs to match applicants with the most suitable scholarship based on their individual backgrounds and qualifications. A number of measures, including accuracy, precision, recall, and F1-score, were used to assess the performance of the models. Results revealed Logistic Regression as the best-performing model in terms of overall accuracy and balance between precision and recall. Moreover, the Support Vector Machine, Naive Bayes, and Random Forest models demonstrated competitive performance, while the Multilayer Perceptron exhibited the lowest performance among others. This study developed machine learning models to improve scholarship selection, testing algorithms like Naïve Bayes, Random Forest, Logistic Regression, SVM, and MLP. Logistic Regression achieved the best overall performance, offering a more accurate and balanced approach to matching applicants with suitable scholarships.

2024-01-01T00:00:00Z Abong, Edgardo Jr S. Janducayan, Karelle Teyle A. Lima, Jomer Mae M. Aborde, Meljohn V. https://repository.umindanao.edu.ph/handle/123456789/2320 2026-05-07T11:13:23Z 2024-01-01T00:00:00Z

An optimized mask R-CNN with bag-of-visual words and fast+surf algorithm in sharp object instance segmentation for x-ray security Abong, Edgardo Jr S.; Janducayan, Karelle Teyle A.; Lima, Jomer Mae M.; Aborde, Meljohn V. Automated security X-ray analysis is highly desired for efficiently inspecting sharp objects. The research formulated an optimized approach for sharp object detection using a Mask R-CNN architecture. The dataset used during the training phase consists of 238 balanced raw images extracted from GitHub named OPIXray. The researchers utilized recent advances in computer vision algorithms, including the Bag-of-Words and Fast+Surf feature extraction techniques, to improve the accuracy and reliability of object deletion. The research demonstrated that the optimized versions of the classification and object detection models have significantly improved accuracy for most categories, with a 5% improvement for the clear category and a 3% improvement for both the scissor and straight knife detection. This study developed a Mask R-CNN–based model to detect sharp objects in X-ray images, using the OPIXray dataset. The optimized approach improved detection accuracy, with notable gains in identifying clear images (+5%) and sharp objects like scissors and knives (+3%).

2024-01-01T00:00:00Z Obra, Mark Japhet D. Murcia, John Vianne B. https://repository.umindanao.edu.ph/handle/123456789/2314 2026-05-07T11:12:15Z 2024-01-01T00:00:00Z

Predicting physical activity engagement among college students : a logistic regression analysis of lifestyle influences Obra, Mark Japhet D.; Murcia, John Vianne B. This research aimed to create a model that could predict how likely college students are to participate in physical activities (PA) using a method called logistic regression analysis. The study collected data from 1,118 students at the largest private university in Mindanao, Philippines, through a survey called the physical activity engagement survey (PAES). An analysis of the students' backgrounds found that most of them were single, did not have jobs, were female, and did not have children. The study also looked at the students' lifestyles and found many did not have a family history of illness but did consume a lot of junk food and alcohol and did not often walk from home to school. The analysis using logistic regression found important factors that could predict if students would take part in PA. It was discovered that male students were more likely to be active than female students. Surprisingly, students studying degrees that required physical effort and those who had jobs were less likely to be active. On the other hand, students who had a strong interest in PA and knew its benefits were more likely to be active. The model was able to correctly identify whether 72.9% of the students were active or not. Additionally, it was noted that students often ate out, consuming a lot of burgers, fried foods, sweets, and sugary drinks. The study also looked at how students viewed their physical education (PE) classes, their own fitness levels, and how effective they thought their PE teachers and facilities were. Overall, students had a positive view but also pointed out some areas that could be better. This detailed analysis shows that many different factors, like background, lifestyle, and perceptions, play a role in whether college students are likely to engage in PA. The findings suggest that efforts to encourage PA among students should be tailored to address these diverse factors. This study used logistic regression to predict college students’ participation in physical activities, analyzing data from 1,118 students in Mindanao. The model achieved 72.9% accuracy and revealed that gender, lifestyle, and perceptions strongly influence physical activity engagement.

2024-01-01T00:00:00Z