Research on COVID-19 prediction by Polytechnic students published in 1st-class international journal
Students Nguyen Ho Quang, Vo Quang Nghia and Nguyen Quang Duc – Photo: THY HUYEN
The article titled: BeCaked – An Explainable Artificial Intelligence Model For COVID-19 Forecasting (BeCaked – An Explainable Artificial Intelligence Model for COVID-19 Forecasting) was approved for publication on Journal Scientific Reports rated Q1 by Springer Nature Publishing.
The author group of this study consists of 3 students: Nguyen Quang Duc, Nguyen Ho Quang (Faculty of Computer Science and Engineering, University of Science and Technology) and Vo Quang Nghia (Faculty of Medicine – Vietnam National University, Ho Chi Minh City), under the guidance of lecturers from the University of Science and Technology and the University of Medicine and Pharmacy in Ho Chi Minh City.
Nguyen Quang Duc has developed a machine learning model capable of directly integrating the SIRD dynamics system and achieved some pretty good results. Then, this model is taken to competitions, conferences, seminars to get feedback and improvement. This is also the project that won the first prize in the field of biotechnology – biomedical science student competition Eureka 2020.
After that, the system was gradually upgraded and transferred to the Department of Health of Ho Chi Minh City after achieving 98% accuracy in predicting the number of cases, infections and deaths worldwide.
Not stopping there, the group continued to improve and research, and with the help of instructors, the group wrote the model into an article and submitted it to the journal. Scientific Reports and was approved for posting on April 26.
Research by the BeCaked system is done by collecting actual data on the epidemic situation (number of new infections, recoveries, deaths) that are publicly available from reliable sources. From there, the model will apply advanced AI techniques based on deep learning techniques to infer the intrinsic parameters of the epidemic, such as direct and indirect infection frequencies.
In addition, the inference of parameters also enables BeCaked to help experts and policymakers explain the causes of the epidemic situation and devise appropriate response measures.
Experimental results with real data show that BeCaked has very high accuracy, especially compared with similar works published in the world. Currently, BeCaked’s forecast results are being deployed at the website: http://cse.hcmut.edu.vn/BeCaked.
Assoc. Prof. Dr. Quan Thanh Tho – Vice Dean of Computer Science Department, University of Science and Technology, instructor for the group – said: “The group’s research involved a freshman student, currently when the article is published. If you approve the publication, I will go to second year. In addition, the research does not stop at a certain point, but always improves to improve the quality of the work.”
Scientific Reports was established in 2011 as a multidisciplinary journal specializing in publishing articles on natural sciences, ranked Q1 by Springer Nature Publishing House and considered as a competitor to the journal. Plos One famous (Plos One is one of the world’s leading multidisciplinary journals).
Scientific Reports is in the top 10 of the most prestigious multidisciplinary journals evaluated by the SCImago Foundation based on its contributions to the scientific community.
In 2016, Scientific Reports The journal has the largest number of published articles in the world. In addition, Springer Nature Publishing is one of the world’s leading publishers of the natural sciences and medicine.
at Blogtuan.info – Source: tuoitre.vn – Read the original article here
