Shaneth C. Ambat
12 Publications
Scopus ID: 105007536684
2024 International Conference on IT Innovation and Knowledge Discovery (ITIKD), (2025), pp. 1-6
Conference Paper | Published: January 1, 2025
Abstract
Part of speech tagging (POS) is crucial in natural language processing, identifying the grammatical categories of words in sentences. This research highlights the lack of focus on POS tagging for Asian languages, particularly Cebuano. Inadequate research on Cebuano religious text has hindered linguistic documentation and understanding its grammar and vocabulary. This study introduces a Parts-of-Speech Tagging for Cebuano utilizing a Modified Viterbi Algorithm. The researchers also apply a method for handling unfamiliar words. Results indicate that the algorithm performs exceptionally well on a religious text corpus comprising 50,000 datasets, achieving an accuracy of93%,precision of90%, recall of 90. 52%, and an F1-score of92%. These results highlight the algorithm's effectiveness in tackling language challenges within specific genres. Furthermore, the research supports the Sustainable Development Goals (SDGs) by promoting linguistic diversity and advancing inclusive language technologies. The study also provides valuable insights into Cebuano's linguistic characteristics and grammatical structures, laying a solid foundation for future research in natural language processing.
Scopus ID: 105007560277
2024 International Conference on IT Innovation and Knowledge Discovery (ITIKD), (2025), pp. 1-6
Conference Paper | Published: January 1, 2025
Abstract
In this study, we use various RNN architectures namely, RNN, Bi-LSTM, and GRU — alongside BERT to analyze sentiment across university departments. Our aim is a comparative analysis of these models in sentiment classification within education. We collected and pre-processed textual data from multiple departments for balanced training and validation. Results showed that traditional RNNs achieved 90% accuracy, Bi-LSTM 93%, and GRU 89%. BERT, leveraging its Transformer architecture, outperformed with 94% accuracy. These findings highlight the superiority of BERT in capturing complex language patterns for sentiment analysis. This study underscores the potential of advanced neural network architectures to gain insights into departmental sentiments, informing policy decisions and educational strategies. Aligning with sustainable development goals in education, we aim to use AI models to develop effective, inclusive, and responsive educational strategies, enhancing quality and accessibility.
Scopus ID: 105000268427
TENCON 2024 - 2024 IEEE Region 10 Conference (TENCON), (2024), pp. 940-943
Conference Paper | Published: January 1, 2024
Abstract
Part of speech tagging (POS) is crucial in natural language processing, identifying the grammatical categories of words in sentences. This research highlights the lack of focus on POS tagging for Asian languages, particularly Cebuano. Limited research on Cebuano has hindered linguistic documentation and understanding of its grammar and vocabulary. This study introduces a Cebuano POS tagger using the Hidden Markov Model (HMM) to improve Cebuano text processing. The researchers also propose a method for handling unfamiliar words. Results show the algorithm performs well on a news text corpus of 25,000 datasets, with an accuracy of 84 %, precision of 80%, recall of 81.52%, and F1-score of 82%. These outcomes demonstrate the algorithm's effectiveness in addressing language challenges in specific genres. Additionally, the research contributes to the Sustainable Development Goals (SDGs) by promoting linguistic diversity and fostering inclusive language technologies. The study provides insights into Cebuano's linguistic traits and grammatical structures, offering a foundation for further research in natural language processing.
Scopus ID: 85199910971
2023 IEEE 15th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment, and Management (HNICEM), (2023), pp. 1-4
Conference Paper | Published: January 1, 2023
Abstract
This research study employed machine learning algorithm in This research study employed a machine learning algorithm in predicting student academic performance in the Data Structures and Algorithm (DSA) course which is based on student lifestyle to analyze the factors that affect the high or low performance result. A total number of 251 Bachelor of Science in Computer Science (BSCS) students participated in the study where 207 or 82% were male and 44 or 18% were female. A oneshot case study was conducted that led to data collection through the administration of an online survey on former enrollees of the said course. The dataset was extracted with 43 features and was analyzed using Python on Jupyter Notebook. Randomly selected 70% of these, 176 observations, are used to train the classifier models. The remaining 30%, 75 observations, were used as the test data. In order to classify academic performance students, eight machine learning algorithms were applied based on random forest (RF), decision tree (DT), support vector machines (SVM), K-nearest neighbors (KNN), logistic regression (LR), Gaussian Naive Bayes (GNB), stochastic gradient descent (SGD), and perceptron. Although SGD and Perceptron classifier models show comparably low classification performances, both random forest and decision tree classifiers provided the highest metric performance. The study indicated that the lifestyles of students contributed to whether the student performance became high or low in their grade performance.
Scopus ID: 85159403155
2022 IEEE 14th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment, and Management (HNICEM), (2022), pp. 1-5
Conference Paper | Published: January 1, 2022
Abstract
Computer programming is now one of the most critical competencies taught in computer courses. [1]. Students require any assistance they can get when learning programming in order to acquire the necessary abilities to excel in the field of computing [2]. This paper aims to investigate the C compiler error logs of Computer Science freshmen students. A prototype was developed and pilot-tested to obtain C source code snippets focusing on assignment statements. The dataset consisting of 1013 logs were extracted from the initial prototype then followed the data science approach of [3] for pre-processing. A Person correlational analysis was conducted on eight features to investigate the relationship between all variables in the dataset. Results of the study show that there is a strong relationship between wrong expression and operator (0.806), wrong expression and numeric value (0.794), operator and numeric value (0.663). Implications of this study is also helpful to computing instructors to improvise the delivery of their teaching pedagogy.
Scopus ID: 85123285806
Lecture Notes in Networks and Systems, (2022), pp. 575-583
Book Chapter | Published: January 1, 2022
Abstract
Optical character recognition (OCR) is the process of extracting the characters from a digital image. The concept behind OCR is to acquire a text in a video or image formats and extract the characters from that image and present it to the user in an editable format. In this study, a convolutional neural network (CNN) is applied, which is a mathematical representation of the functionality of the human brain, using back-propagation algorithm with test case files of English alphabets and numbers. The purpose of this study is to test systems capable of recognizing vehicle plate number English alphabets and numbers with different fonts, and to be familiar with CNN and digital image processing applied for character recognition. Scientific journals and reports were used to research the relevant information required for the thesis project. The chosen software was then trained and tested with both computer and video output files. The tests revealed that the OCR software can recognize both vehicular plate and computer alphabets and learns to do it better with each iteration. The study shows that although the system needs more training for vehicular plate characters than computerized fonts, and the use of CNN in OCR is of great benefit and allows for quicker and better character recognition.
Scopus ID: 85123284494
Lecture Notes in Networks and Systems, (2022), pp. 691-697
Book Chapter | Published: January 1, 2022
Abstract
E-Aid is a study that aims to develop an application based on the convolutional neural network algorithm. The central idea for the creation of E-Aid is to provide a mobile application which offers more advanced capabilities and leads to a strong emergence for the medical health applications in the market. The reliability for the usage of CNN as an algorithm produces positive results which is essential for this study. The researchers trained CNN model that will be used later on during the execution of the CCN algorithm, and this CNN model must be able to identify 4 types of open wounds (laceration, puncture, abrasion and avulsion) and 4 types of skin burns (1st-, 2nd-, 3rd- and 4th-degree burn) and also must be able to classify it whether the wound is infected or not infected. The researchers tested the accuracy of the CNN model before sending to our respondents. The researchers tested the accuracy by getting a random image of open wounds and skin burns in the Internet and run it on the E-Aid app. After the researchers finish testing the accuracy of the app, they distributed the app to their respondents to test furthermore the accuracy and reliability of the app. The researchers’ respondents are composed of 6 medical professionals (doctors/nurses), 5 IT/CS professionals and 14 students (in the field of medicine and computer studies).
Scopus ID: 85134878533
2022 IEEE World AI IoT Congress (AIIoT), (2022), pp. 111-117
Conference Paper | Published: January 1, 2022
Abstract
Efficient and reliable medical image analysis is indispensable in modern healthcare settings. The conventional approaches in diagnostics and evaluations from a mere picture are complex. It often leads to subjectivity due to experts' various experiences and expertise. Using convolutional neural networks, we proposed an end-to-end pipeline for automatic Cobb angle measurement to pinpoint scoliosis severity. Our results show that the Residual U-Net architecture provides vertebrae average segmentation accuracy of 92.95% based on Dice and Jaccard similarity coefficients. Furthermore, a comparative benchmark between physician's measurement and our machine-driven approach produces an acceptable mean deviation of 1.57 degrees and a T-test p-value of 0.9028, indicating no significant difference. This study has the potential to help doctors in prompt scoliosis magnitude assessments.
Scopus ID: 85159384712
2022 IEEE 14th International Conference on Humanoid, Nanotechnology, Information Technology, Communication and Control, Environment, and Management (HNICEM), (2022), pp. 1-6
Conference Paper | Published: January 1, 2022
Abstract
Computer programming is one of the fundamental skills in the field of computing [1]. In a computing class, students are expected to learn the skills rather than remembering materials only. This study aims to develop a constraint-based student model (CBM) by analyzing the computing students' C compilation error logs. The proposed modified CBM will be used as input to develop a user behavior of an ongoing study for an intelligent tutoring system. The prototype was developed to obtain compilation error logs from the selected students, it contains five (5) C programming questions that focus on assignment statements. The prototype of the study was pilot tested on two (2) online programming classes with a total of thirty-one (31) freshman college students composed of nine (9) BSCS and twenty-two (22) BSIT participants with a mean age of 18.68, where nineteen (19) or 61.3% are males and twelve (12) or 38.7% are females. The study uses convenience sampling to determine the total number of student participants. The dataset was extracted from the prototype and feature identification was performed on one thousand thirteen (1013) C programming logs which resulted to obtain eight (8) error types. The paper of Khodeir, Wanas, & Elazhary (2018) [2] and Karaci (2018) [3] on constraint-based modeling was reviewed to develop a proposed constraint-based model in the context of C programming focusing on assignment statements. By mapping a student error on the suggested constraint relevance (Cr) and constraint satisfaction, the database for constraints was finished (Cs).
Scopus ID: 85132193982
2022 10th International Conference on Information and Education Technology (ICIET), (2022), pp. 390-397
Conference Paper | Published: January 1, 2022
Abstract
Forecasting the number of incidences of medical cases is important in planning institutional health program strategies to draft intervention and allocate resources. The utilization of advancements in computing and the use of massive health data create possibilities for the generation of tools in a recommender system. This study focused on medical cases forecasting using exponential smoothing model for the development of resource allocation recommender system. Different data pre-processing techniques were used such as imputation and data cleaning in the historical dataset. To determine which set of alpha values can be considered and be used in the development of online resource allocation recommender system for Mabalacat City Health Unit, the mean absolute percent error and mean absolute deviation were used. Exponential smoothing with an alpha value of 0.9 and 0.3 have high forecasted values than that of Exponential smoothing using 0.1, 0.5 and 0.7 respectively.