1 January 2015 in Procedia Computer Science
Brain Computer Interfaces, abbreviated as BCI, is a technology which allows users to take action in computer by focusing on which action the user wants to do. BCI processes brainwave which recorded by means of electroencephalography so it is can be known to computers. Inside the brain, nerve impulses is passed across corresponding nerves to command body part into action, as a result, the respective body part responded with the action that brain commanded. This chain of events can be harnessed by BCI so that brainwaves can be acquired. Using a proper classification method, brainwaves data can be used as a digital command to computer, eliminating the need of actual real-world action to act in a computer. One of the BCI's interesting topic is Motor Imagery, a topic which deeply examine brain activity when imagining motoric activity such as moving left hand. Brain activity is recorded by electroencephalography so that BCI can use the data as a reference to translate brain activity into actual motor activity or as computer trigger. Emotiv EPOC is one of the common BCI distributed to public users. The purpose of this research is to determine how BCI can identify and distinguish human's brainwave when performing different activities. Emotiv EPOC is used in this research with a purpose of whether Emotiv EPOC can be used for Motor Imagery. A simple method is used in this research, using Graz BCI scenario provided in OpenVibe installation bundle and then assess the test results with 5seconds timeframe. A total of 6 scenarios is used for testing purpose. The success rates of those 6 scenarios are: scenario 1 with 76,67% success rate, scenario 2 with 91,67% success rate, scenario 3 with 28,33% success rate, scenario 4 with 13,33% success rate, scenario 5 with 60% success rate, and scenario 6 with 76,67% success rate. The result from this research indicates that Emotiv EPOC can be a possible option but not recommended for implementing motor imagery application
1 September 2020 in Health Policy and Technology
1 June 2023 in International Journal of Electrical and Computer Engineering (IJECE)
Click counts are related to the amount of money that online advertisers paid to news sites. Such business models forced some news sites to employ a dirty trick of click-baiting, i.e., using hyperbolic and interesting words, sometimes unfinished sentences in a headline to purposefully tease the readers. Some Indonesian online news sites also joined the party of clickbait, which indirectly degrade other established news sites' credibility. A neural network with a pre-trained language model multilingual bidirectional encoder representations from transformers (BERT) that acted as an embedding layer is then combined with a 100 node-hidden layer and topped with a sigmoid classifier was trained to detect clickbait headlines. With a total of 6,632 headlines as a training dataset, the classifier performed remarkably well. Evaluated with 5-fold cross-validation, it has an accuracy score of 0.914, an F1-score of 0.914, a precision score of 0.916, and a receiver operating characteristic-area under curve (ROC-AUC) of 0.92. The usage of multilingual BERT in the Indonesian text classification task was tested and is possible to be enhanced further. Future possibilities, societal impact, and limitations of clickbait detection are discussed.
1 December 2022 in International Journal of Electrical and Computer Engineering (IJECE)
We proposed an approach of retailer segmentation using a hybrid swarm intelligence algorithm and recency frequency monetary (RFM)-location model to develop a tailored marketing strategy for a pharmacy industry distribution company. We used sales data and plug it into MATLAB to implement ant clustering algorithm and K-means, then the results were analyzed using RFM-location model to calculate each clusters’ customer lifetime value (CLV). The algorithm generated 13 clusters of retailers based on provided data with a total of 1,138 retailers. Then, using RFM-location, some clusters were combined due to identical characteristics, the final clusters amounted to 8 clusters with unique characteristics. The findings can inform the decision-making process of the company, especially in prioritizing retailer segments and developing a tailored marketing strategy. We used a hybrid algorithm by leveraging the advantage of swarm intelligence and the power of K-means to cluster the retailers, then we further added value to the generated clusters by analyzing it using RFM-location model and CLV. However, location as a variable may not be relevant in smaller countries or developed countries, because the shipping cost may not be a problem.
1 December 2022 in IAES International Journal of Artificial Intelligence (IJ-AI)
This study attempted to deploy a high performing natural language processing model which specifically trained on flagging clickbait Indonesian news headline. The deployed model is accessible from any internet-connected device because it implements representational state transfer application programming interface (RESTful API). The application is useful to avoid clickbait news which often solely purposed to rack money but not delivering trustworthy news. With many online news outlets adopting the click-based advertising, clickbait headline become ubiquitous. Thus, newsworthy articles often cluttered with clickbait news. Leveraging state-of-the-art bidirectional encoder representation from transformers (BERT), a lightweight web application is developed. This study offloaded the computing resources needed to train the model on a separate instance of virtual server and then deployed the trained model on the cloud, while the client-side application only needs to send a request to the API and the cloud server will handle the rest, often known as three-layer architecture. This study designed and developed a web-based application to detect clickbait in Indonesian using IndoBERT as its language model. The application usage and potentials were discussed. The source code and running application are available for public with a performance of mean receiver operating characteristic-area under the curve (ROC-AUC) of 89%.
1 August 2022 in Bulletin of Electrical Engineering and Informatics
COVID-19 pandemic changed how society behaves. Travel and social restrictions, commonly associated with the term lockdown became popular and ubiquitous. Given the rise of gig economy and mobile app delivery in the past several years, combined with lockdowns during the pandemic, and the application of telemedicine becomes essential. Halodoc is one of the popular telemedicine applications in Indonesia, having several useful features such as text-based doctor consultation and prescription drug order-delivery, and Halodoc is easily preferred by many. This article explored the motivation behind using Halodoc as the preferred method of getting health service during the pandemic, behind the perceived usefulness and perceived ease of use of the application, we found that fear of missing out (FOMO) has an indirect role in the application adoption in society, especially during lockdowns, where social interaction is limited to social media and other internet-based platforms. The reason why FOMO can be an important factor in technology adoption and how advertisers should explore FOMO is further discussed.