I'm currently a chatbot developer in EastWest Banking Corporation. A bot to allow credit card application
and provide real-time notifications.
An avenue for customer feedback and using NLP to give the best response on their specific concerns or
queries.
Researching other potential avenues that a chatbot can help to improve both internal and customer-related
business operations.
Personal Information
Origin:
Samal, Island Garden City of Samal
Education:
BS in Computer Science, University of the Philippines Mindanao
Nickname:
Al
Skills
Conversational AI
Developed ESTA, EastWest System Technical Assistant.
Esta serves as a digital tool to help customers apply for credit card, check their application status,
and submit application requirements.
Certain promo codes and referral links could allow Esta for auto loans application, personal loan and
other products by consumer lending division.
Esta has two main channels, on Facebook Messenger the most popular communication platform used by
Filipinos, and a secure webchat
which information and documents are uploaded as the company values information and data privacy of their
customers.
Esta also integrates LUIS (for natural language processing) and CustomVision (for image detection).
Esta can respond to greetings and spontaenous follow up questions made possible by LUIS that was setup.
The CustomVision feature allows Esta to distinguish valid government ID from not to avoid customers
uploading erroneous requirements.
Its a continuous learning on how to improve ESTA on providing customers' needs.
Bioinformatics
6-weeks training and internship program for bioinformatics at the Philippine Genome Center
Research project focused on Molecular Dynamics Simulation and Modelling of integrins.
Supervised by Mr. Albao from the project of Dr. Bascos of Protein Structure and Immunology Laboratory of
National Institute of Molecular Biology and Biotechnology.
Poster presentation at 2019 International Union of Biochemistry and Molecular Biology and Philippine
Society for Biochemistry and Molecular Biology
Sentiment Analysis
Filipino Tweets Sentiment Analysis using Convolutional Neural Network with Distant Supervision
ABSTRACT
Philippines being labelled as the social media capital of the world means that Filipinos spend their time
in the internet for recreation and entertainment purposes.
Social media is the main avenue of people to share stories and express their sentiments in various topics
ranging from politics to media personalities.
All these sentiments suggests that there is a huge amount of Filipino sentiment data out there waiting to
be converted into valuable information.
Current approaches used in natural language processing for Filipino language is the classic machine
learning algorithms such as Support Vector Machine and Naïve Bayes which yielded low accuracy or used
imbalanced data set for training.
In this paper, the researcher explored Convolutional Neural Networks by analyzing sentiments of Filipino
Tweets.
The network performs binary classification and initially labelled their tweets based on their emoticons.
For the architecture to accept tweets, before training, the tweets were converted to vectors using
word2vec.
The network was trained to 7,741 tweets and tested to 1,936 tweets.
The model yielded with accuracy of 74.68% accuracy and was compared to Naïve Bayes that yielded 69.49% by
the same dataset.
This small margin between the classic machine learning algorithm means that there is a potential in
exploring CNN architecture for various natural language processing problems for Filipino languages.
