The Investigation into Deep Learning Classifiers Towards Imbalanced Text Data

Class imbalance is an important classification problem where failure to identify events can be hazardous due to failure of solution preparation or opportune handling. Minorities are mostly more consequential in such cases. It is necessary to know a reliable classifier for imbalanced classes. This study examines several conventional machine learning and deep learning methods to compare the performance of each method on dataset with imbalanced classes. We use COVID-19 online news titles to simulate different class imbalance ratios. The results of our study demonstrate the superiority of the CNN with embedding layer method on a news titles dataset of 16,844 data points towards imbalance ratios of 37%, 30%, 20%, 10%, and 1%. However, CNN with embedding layer showed a noticeable performance degradation at an imbalance ratio of 1%. © 2022 IEEE.

Synonym-based Text Generation in Restructuring Imbalanced Dataset for Deep Learning Models

One of which machine learning data processing problems is imbalanced classes. Imbalanced classes could potentially cause bias towards the majority classes due to the nature of machine learning algorithms that presume that the object cardinality in classes is around similar number. Oversampling or generating new objects in minority class are common approaches for balancing the dataset. In text oversampling method, semantic meaning loses often occur when deep learning algorithms are used. We propose synonym-based text generation for restructuring the imbalanced COVID-19 online-news dataset. Three deep learning models (MLP, CNN, and LSTM) using TF/IDF and word embedding (WE) feature are tested with the original and balanced dataset. The results indicate that the balance condition of the dataset and the use of text representative features affect the performance of the deep learning model. Using balanced data and deep learning models with WE greatly affect the classification significantly higher performances as high as 4%, 5%, and 6% in accuracy, precision, recall, and f1-score, respectively. © 2022 IEEE.

DCPC : Drug Candidates for the Prevention of COVID-1 9 Database

The world immediately studied Coronavirus Disease 2019 (COVID-19) and raced towards fmding the cure and developing an effective treatment. An automated approach is needed to discover drug candidates and provide those data to facilitate clinical trials in saving time and only focusing on the candidates which potentially become the cure for COVID-19. We propose the Drug Candidates for the Prevention of COVID-19 (DCPC) Database. DCPC Database provides a list of candidates of potential drugs for the prevention of COVID-19 based on disease-drug associations which are automatically discovered from biomedical literature. DCPC database is an integrative structural database, which involves a chemical database repository, such as PubChem and DrugBank to ensure that drug compound candidates have a standard representation of compounds. The database provides keyword-chosen categories and a determination of minimum supported articles for search, a list of drug candidates in the sorted table followed by the detail for each candidate, and a download feature. The keyword category consists of three keywords, they are Chinese herbal compounds, Indian medicinal plants, and Indian medicinal plants & diabetic treatment herbs. Each candidate links to an article in the biomedical literature and to a page of the compound structure visualization. DCPC is freely available at https://dcpc.brin.go.id/dcpc/.

Monitoring Indonesian online news for COVID-19 event detection using deep learning

Even though coronavirus disease 2019 (COVID-19) vaccination has been done, preparedness for the possibility of the next outbreak wave is still needed with new mutations and virus variants. A near real-time surveillance system is required to provide the stakeholders, especially the public, to act in a timely response. Due to the hierarchical structure, epidemic reporting is usually slow particularly when passing jurisdictional borders. This condition could lead to time gaps for public awareness of new and emerging events of infectious diseases. Online news is a potential source for COVID-19 monitoring because it reports almost every infectious disease incident globally. However, the news does not report only about COVID-19 events, but also various information related to COVID-19 topics such as the economic impact, health tips, and others. We developed a framework for online news monitoring and applied sentence classification for news titles using deep learning to distinguish between COVID-19 events and non-event news. The classification results showed that the fine-tuned bidirectional encoder representations from transformers (BERT) trained with Bahasa Indonesia achieved the highest performance (accuracy: 95.16%, precision: 94.71%, recall: 94.32%, F1-score: 94.51%). Interestingly, our framework was able to identify news that reports the new COVID strain from the United Kingdom (UK) as an event news, 13 days before the Indonesian officials closed the border for foreigners. © 2023 Institute of Advanced Engineering and Science. All rights reserved.

DCPC: Drug Candidates for the Prevention of COVID-19 Database

The world immediately studied Coronavirus Disease 2019 (COVID-19) and raced towards finding the cure and developing an effective treatment. An automated approach is needed to discover drug candidates and provide those data to facilitate clinical trials in saving time and only focusing on the candidates which potentially become the cure for COVID-19. We propose the Drug Candidates for the Prevention of COVID-19 (DCPC) Database. DCPC Database provides a list of candidates of potential drugs for the prevention of COVID-19 based on disease-drug associations which are automatically discovered from biomedical literature. DCPC database is an integrative structural database, which involves a chemical database repository, such as PubChem and DrugBank to ensure that drug compound candidates have a standard representation of compounds. The database provides keyword-chosen categories and a determination of minimum supported articles for search, a list of drug candidates in the sorted table followed by the detail for each candidate, and a download feature. The keyword category consists of three keywords, they are Chinese herbal compounds, Indian medicinal plants/and Indian medicinal plants & diabetic treatment herbs. Each candidate links to an article in the biomedical literature and to a page of the compound structure visualization. DCPC is freely available at https://dcpc.brin.go.id/dcpc/. © 2022 IEEE.

Development of a low-cost wearable device for Covid-19 self-quarantine monitoring system.

Objectives: The objective of this study is to develop a Bluetooth-based low-cost wearable device for a self-quarantine monitoring system. Study design: The designed wearable device focuses on data transmission via Bluetooth, integration of tracking, tracing, and fencing into a single system, and low energy usage from its battery. Methods: We design a wearable device using smartphone equipped with GPS, a communication module, Bluetooth low energy (BLE) and a high-capacity battery as a solution for low-cost device with excellent efficiency. We divide the designed system into two parts, the client and the server parts. The client parts are wearable device attached to the individual being monitored and the mobile phone as GPS and telecommunications module. Whereas the server parts are user interface, digital map, notification system, and backend database. Then, the whole system was tested in laboratory and field scale. Results: We tested functions of integrated device such as wearable device, mobile applications, and server for laboratory scale test. Then, performing field test with geofencing, communication module, battery, web interface, and resource computing usage. The field test was conducted on a small scale with a limited number of trial patients. We found that the designed wearable device was successfully implemented for both self-quarantine and centralized quarantine requirements. The majority of the components used met the specifications and functioned properly as well. Conclusions: A BLE-enabled wearable device can be used for tracking self-quarantine patients. The laboratory and field scale tests demonstrate that the designed wearable device functions properly and meets the requirements. We anticipate that this low-cost wearable device is effective in limiting Covid-19 virus spread and preventing the formation of a new Covid-19 virus-infected cluster.

Classifying aggravation status of COVID-19 event from short-text using CNN

COVID-19 pandemic is a new precedent that has changed many aspects of human life. With the uncertainty of vaccine availability, stakeholders are required to track the dynamics of COVID-19 events to prepare the necessary response. One sub-task in tracking the dynamics of an event is to identify the aggravation status of the event (i.e., whether an event is worsening or getting better). We experimented with convolutional neural network (CNN) models to classify the status of COVID-19 aggravation status from a short text. CNN without one hot encoding prevailed. Furthermore, we conduct tuning to achieve better performance of CNN. The highest performance was achieved by tuning some of the configuration parameters. As the final result, the model performed at best (accuracy = 87.585% and F1-score = 76%) when using 80 nodes, SGD optimizer, lr = 0.1, and momentum = 0.9.