JIJoseph Marvin Imperial

Joseph Marvin Imperial

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Joseph Marvin Imperial has a total of 16 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Changing Topics for Changing Times: Thematic and Temporal-Based Analysis of the Philippine Senatorial and Midterm Elections

    by Lamar Clarence Cruz, Jessica Nicole Dela Cruz, Shane Francis Maglangit, Mico Magtira, Joseph Marvin Imperial, Ramon Rodriguez

    Elections in the Philippines are one of the most prominent events in the country. It is a time when people choose their leader through the democratic election process. During the electoral campaign, candidates try several strategies to win the election. However, victors emerge are those familiar and have mass-based support. In this paper, we have investigated the perception of netizens in the Philippines for the last two election seasons, 2019 and 2022, based on Twitter data. Thematic analyses were conducted to determine the themes related to winning candidates and create a link on how they may affect their election success. The study found that the top candidates (Cynthia Villar in 2019 and Robinhood Padilla in 2022) mainly received criticism and opposition towards their campaigns from the themes. Moreover, findings suggest that candidates with more negative or positive publicity emerge as winners. Long-time politicians and organized groups endorsed most winners during the 2019 and 2022 elections.

  • Open Access

    WikAnalytics: A Web-based Application for Identifying Linguistic Features of a Text Group Supporting Filipino, English, and Taglish Languages

    by Jomari Valmadrid Ramos, John Michael Ballesta, Andrew Kobe Lee Yam, Moises Kairon Mogol, Ramon Rodriguez, Joseph Marvin Imperial

    Reading is one of the first things humans learn to improve their comprehension, vocabulary, and imagination. Determining what to read for your level will be difficult, but forcing you to understand a high text level is more complex. It can result in not learning at all. The Philippines has two official languages, which are English and Filipino. Filipinos tend to engage with text written in both languages. Although the Philippines has two official languages, not all individuals are bilingual and only focus on one language, especially those who do not have the privilege to study in school. Thus, having bilingual text to read might be challenging to understand its context. A system that can analyze and recognize either English, Filipino, or Taglish languages while giving the readability index of the given text could help readers determine the readability level they want to read. Wikanalytics, a web application text analysis tool, was developed in this paper. The application aims to provide analysis of text files while calculating its readability index. The Agile Software Development Method was utilized for developing the web application. The system handles text documents written in English, Filipino, or Taglish Languages. The extracted features of an analyzed text are Readability Index, Traditional Features, Lexical Features, Syllable Patterns, Sentences and Paragraphs, and Token Ratios. The accuracy of the web application tool in identifying the language is 97.5% in English, 96% for Filipino, and 96.89% for Taglish. The runtime is tested for English, Filipino, and Taglish files, and the longest runtime out of the three languages mentioned is the Taglish file. Having two different languages in the file affected the runtime of the application's Lexical and Token ratio feature.

  • Open Access

    A BERT-based Hate Speech Classifier from Transcribed Online Short-Form Videos

    by Rommel Hernandez Urbano Jr., Jeffrey Uy Ajero, Angelic Legaspi Angeles, Maria Nikki Hacar Quintos, Joseph Marvin Regalado Imperial, Ramon Llabanes Rodriguez

    With the rise of human-centric technologies such as social media platforms, the amount of hate also continues to grow proportionally with the increasing number of users worldwide. TikTok is one of the most-used social media platforms due to its feature that allows users to express themselves via creating and sharing short-form videos based on any desired topic and content. In addition, it has also become a platform for political discourse and mudslinging as users can freely express an opinion and indirectly debate with random people online. In this study, we propose the use of BERT, a complex bidirectional transformer-based model, for the task of automatic hate speech detection from speech transcribed from Tagalog TikTok videos. Results of our experiments show that a BERT-based hate speech classifier scores 61% F1. We also extended the task beyond several algorithms such as LSTM, Naïve Bayes, and Decision Tree and found out that traditional methods such as a simple Bernoulli Naïve Bayes approach remain at par with the BERT model.


  • Open Access

    Diverse Linguistic Features for Assessing Reading Difficulty of Educational Filipino Texts

    by Joseph Marvin Imperial, Ethel Ong

    In order to ensure quality and effective learning, fluency, and comprehension, the proper identification of the difficulty levels of reading materials should be observed. In this paper, we describe the development of automatic machine learning-based readability assessment models for educational Filipino texts using the most diverse set of linguistic features for the language. Results show that using a Random Forest model obtained a high performance of 62.7% in terms of accuracy, and 66.1% when using the optimal combination of feature sets consisting of traditional and syllable pattern-based predictors.

  • Motif search using Gibbs sampling: Notes on effectiveness in a distributed environment

    by Joseph Marvin Imperial, Czeritonnie Gail Ya-On, Gregory Cu

    Motif search is a common problem in bioinformatics where unique DNA sequences (motifs) of a specific length inscribed in long strands signify binding sites for transcription factors. In this paper, we present some important notes on the implementation of motif search using Gibbs sampling algorithm in a distributed computing environment by analyzing visualization on speed and motif scoring of various distributed implementations. For the DNA sequences data, we used an open-source mouse genome fragments with lengths 250, 500, and 1000. We built upon our previous studies (Perera and Ragel, 2013; Chen and Jiang, 2006) by integrating a distributed environment of the motif search workloads (jobs) across 16 CPU cores contained on 2 computer nodes instead of the traditional way of parallelizing on a single computing device with multicore CPUs. Results show that using saving the DNA sequences in list and adding as a parameter argument obtained the fastest execution time compared to implementations by sending file dependencies and in-memory processing.