Dr. Ayu Memaparkan Pengembangan COVID-19 Social Media Monitoring dan Chatbot Triase untuk Rumah Sakit

BANDUNG, itb.ac.id - Humans, as social beings, engage in one fundamental activity: verbal and nonverbal (written) communication. The internet's presence in our daily lives has made communication easier, allowing us to receive information or news quickly, especially during the COVID-19 pandemic that has been occurring in Indonesia since March 2020.

One of the technologies that processes text or sound written in human language is also called Natural Language Processing (NLP). The ITB Artificial Intelligence Center held a webinar series 3 entitled "AI-NLP for Communication During the Pandemic" presented by lecturers from the Informatics Engineering Study Program, School of Electrical Engineering and Informatics (STEI) ITB. Dr. Eng. Ayu Purwarianti, S.T., M.T., on June 5, 2020 on the ITB Artificial Intelligence Center YouTube channel.

Briefly, Ayu explained that the NLP process includes input, processing, and output. When a person speaks, speech recognition recognizes the voice and then processes it through NLP. Similarly, with documents, website, social media. In the process, NLP has 3 main components, namely natural language understanding (converting text or voice into a form that can be understood by the computer), natural language generation (outputting text in natural language like responses .), or text analytics (processing data with output in the form of dashboard or scoring).

“The benefit of NLP technology is to improve user experience in terms of spelling checker, word or text recommendation, become automate support (berupa .), and become search engine And text translation which we often use," explained Ayu.

In her webinar, Ayu discussed . and media monitoring. Ayu explained that the development . done because people are used to using it chat application (WhatsApp, LINE, Telegram, and others). “Position . "It is very strategic and has many advantages, such as 24/7 availability, and is useful for agencies to disseminate information (promotional)," he said.

He explained that currently . still has several weaknesses such as the ability to understand complex statements, the need for a large database of question-answer pairs, and less natural interactions. .This study proposes several solutions to overcome these weaknesses.

For the first weakness, Ayu continued, the proposed solution is with an approach hybrid, namely when . can automatically redirect questions to a human agent when . realized that he did not have knowledge on user questions. For the second weakness, the proposed solution is with a “learning” mechanism from human agents in the approach hybrid and database input in the form of documents, not a list of questions and answers. As for the third weakness, the proposed solution is to add the capability . with speech recognition technology and the addition of a word normalization module to handle slank words.

ITB's Artificial Intelligence Center itself has developed Chatbot Frequently Asked Questions (FAQs) regarding COVID-19 by providing accurate and valid information about the development of COVID-19 in Indonesia and globally. In addition, ITB is collaborating with the Faculty of Medicine, University of Indonesia, to develop chatbot triase COVID-19. Chatbot This is useful as an initial health check for patients in hospitals so that it can minimize medical staff exposure to COVID-19 from patients. Chatbot will classify patient emergencies into mild, moderate and severe groups so that they can be handled directly according to their group.

ITB's second research is COVID-19 Social Media Monitoring Accessible at https://covid19-socmed.id/. This platform aims to monitor the development of COVID-19 issues from social media (Twitter and Instagram) through analysis using related keywords, sentiment, and hoax classifications circulating in Indonesia. The platform's NLP engine, a text analytics engine, has three main components: concept extraction, sentiment classification. hoax classificationThe output from this platform is data presented in the form of a dashboard.

By using certain keywords, COVID-19 Social Media Monitoring can determine the conversation trends of Indonesian society. "We analyze the use of keyword which was frequently used by Indonesians before and during the implementation of Large-Scale Social Restrictions (PSBB). After that, by using sentiment analysis "Then, we can determine whether the topic has a negative, positive, or neutral response based on public response. Furthermore, we can also determine the probability level of hoaxes on a topic," Ayu explained.

Other ITB research is Hoax Classification. Hoax classification requires a database of hoaxes and facts as the initial basis for checking whether a text can be classified as a hoax or a fact. Compared to NLP for English, Ayu explained that the challenge in developing NLP for Indonesian is grammar And mixing language in a single sentence. "Indonesians are accustomed to using more than one language when communicating, including within a single sentence. Furthermore, grammatically, we can understand the meaning of a statement written out of SPOK order," he said.

Reporter: Billy Akbar Prabowo (Metallurgical Engineering, 2016). This news was previously featured on the ITB website.