Academic staff and students from the College of Information and Computing Sciences (CICS) showcased their research in the fields of Artificial Intelligence and Machine Learning at the 10th International Conference on Smart Computing and Communication (ICSCC 2024), held in UNDIKNAS University, Bali, Indonesia in July 2024.
Assoc. Prof. Donata Acula, Ph.D., presented her team鈥檚 paper titled “Sentiment Analysis on Filipino COVID-19 Tweets During the Pandemic Period Using Modified K-Means and Ensemble Model with Supervised Machine Learning Algorithms as Base Estimators.” This research work focused on sentiment analysis using Natural Language Processing (NLP) methods in tweets on COVID-19 that were written in Taglish, English, and Filipino. It employed clustering algorithms, specifically Modified K-Means with Principal Component Analysis (PCA), Percentile Method, and Weighted Average, and explored various classifier algorithms, including the Combination Ensemble Model and Base Estimators like Complement Naive Bayes, Support Vector Machine (SVM), Logistic Regression, and Random Forest. The Modified K-Means approach categorized the dataset into six clusters, achieving Silhouette Coefficient (0.4276), Calinski-Harabasz (15334.6503), and Davies-Bouldin scores (0.8361).
The Combination Ensemble Model produced impressive results, with mean accuracies ranging from 56.48% to 96.13% based on cluster performance. Statistical tests, such as the paired t-test, revealed that while the modified K-Means method did not significantly outperform the weighted average approach, the Combination Ensemble Model marked a significant improvement over Multinomial Naive Bayes.
The paper was co-authored by Carlo Danilo V. Sindayen, Fernando Luis C. Gaite, and Francis Thomas G. Rogando, all BS Computer Science alumni from the Data Science Track. Their collaborative efforts and hard work were evident in the results of this study, which provides valuable insights into sentiment analysis in the context of the COVID-19 pandemic. Acula received the Best Presenter Award under the Artificial Intelligence and Machine Learning track during the closing and awarding ceremony of the event.
At the same conference, another research paper under the mentorship of Acula was presented. “Mental State Recognition Through Speech Analysis Using Bidirectional Long Short-Term Memory Network and Convolutional Neural Network Hybrid Model” was showcased by Axl Julian B. Erquiza, alumni of BSCS Batch 2024 under the Data Science Track. This study explored the effectiveness of deep learning architectures, specifically Convolutional Neural Networks (CNNs) and Bidirectional Long Short-Term Memory (Bi-LSTM) networks, in recognizing mental states through speech analysis. Using the DAIC-WoZ audio dataset, the team classified mental states across five levels, experimenting with three CNN architectures: ResNet50, VGGNet16, and InceptionV3, and their corresponding Bi-LSTM hybrids. The study revealed varied results, showing that ResNet50 with Bi-LSTM improved in precision but did not significantly enhance training accuracy, VGGNet16 maintained stable performance, and InceptionV3 gained little benefit from the Bi-LSTM integration.
This research contributes significantly to the field of speech-based mental health assessment by highlighting the complexities of optimizing deep learning models for such applications. The co-authors of the paper were Russel Janzen E. Mamaclay and Elisha Francia L. Platon, all from the CICS Computer Science Department.
Both research outputs are now available via IEEE Explore and indexed in Scopus.
In addition to her paper presentation, Acula served as the IEEE Young Professional Meet-Up Event moderator. This event, which was hosted in conjunction with ICSCC 2024, was designed to expose the IEEE Young Professional community to graduating students and early-career professionals who were interested in learning more about AI applications. The gathering offered a venue for networking with people in the government, business, and academic sectors, encouraging partnerships and career advancement.
CICS’s commitment to knowledge sharing and cutting-edge research is highlighted by the faculty and student participation in the ICSCC 2024, which highlights the college’s role as a major collaborator. Acula served as one of the Technical Program Committee (TPC) members and was the conference鈥檚 Chair of External Communication.
Other CICS academic staff such as Dr. Eugenia R. Zhuo (IT Department), Dr. Janette E. Sideno (IS Department), Dr. Maricel A.Balais (IT Department), Dr. Mylene J. Domingo (IT Department) and Dr. Noel E. Estrella (IT Department) were among the academic staff who made additional contributions as members of the TPC.
