Computational Research Progress in Applied Science & Engineering CRPASE Vol. 06(02), 114-120, June 2020 ISSN 2423-4591 Forecasting of COVID-19 Confirmed Cases in Vietnam Using Fuzzy Time Series Model Combined with Particle Swarm Optimization Nghiem Van Tinh  Thai Nguyen University of Technology, Thai Nguyen University, Thai Nguyen, Viet Nam Keywords Abstract COVID-19, Forecasting, Fuzzy time series, Fuzzy relationship groups, Particles warm optimization. Since December past year, a novel coronavirus has named COVID-19, which was discovered in Wuhan, China, and has spread to different cities in China as well as to 208 other countries and becomes as a global threat. The spread of virus COVID-19 can put all countries in situation of incapacity of how manage and face. In order to help people to make decisions in dealing this epidemic issue. This study proposes a forecasting model based on fuzzy time series (FTS) and Particle swarm optimization (PSO) for forecasting the number of confirmed cases of COVID-19 in Vietnam. First, the fuzzy relationship groups are utilized to overcome drawbacks of fuzzy relationship matrix in building of fuzzy forecasting model. Second, the PSO algorithm is used to find and adjust the proper number and length of intervals with an intent to achieve the best forecasting accuracy. To verify the effectiveness of the proposed model, a numerical COVID-19 dataset is selected for forecasting process. These forecasting results could be helpful in forecasting future confirmed cases if the spread of the virus did not change very strangely. 1. Introduction COVID-19 has greatly affected over the world after first being reported in Wuhan, China in December-2019. Since then, there has been an exponential growth in the number of such cases around the globe. As of 7th April 2020, the total confirmed cases of 1.345.653 out of which 74.644 have died. It can very well be observed that most countries have reported cases of new Coronavirus. As of now, it has affected 209 countries with COVID-19, including Vietnam. Considering Vietnam, after first emerging in late January 2020, the number remained constant until the beginning of March, when the 17th case returned from London, England, it grew exponentially. As of 7th April, the number of total cases has reached 245 cases. Given the current rate of growth, where can the cases expect to reach in the next 15 days, if no specific precaution is taken. In this context we are really worry of the coming days, and months, everybody now over world, and especially in Vietnam country, ask the question: what's the trend of this virus? under this serious question ,we will trying to give a response by using the forecasting model based on Fuzzy Time Series. The fuzzy time series forecasting models based on fuzzy set theory [1] have been widely applied to diverse fields such as enrolments forecasting [2] - [9], crop productions prediction [10], stock markets [11] and temperature prediction [12]. The fuzzy time series and the corresponding forecast model  Corresponding Author: E-mail address: nghiemvantinh@tnut.edu.vn Received: 10 April 2020; Accepted: 25 May 2020 was introduced by Song and Chissom in 1993. They introduced both the time-invariant fuzzy time series [2] and the time-variant time series [3] model which use the max– min operations to forecast the enrolments of the University of Alabama. Unfortunately, their method has many drawbacks such as huge computation when the fuzzy rule matrix is large and lack of persuasiveness in determining the universe of discourse and the length of intervals. Therefore, Ref. [5] proposed the first-order fuzzy time series model by using simple arithmetic calculations instead of max-min composition operations [3] for better forecasting accuracy. Thereafter, the fuzzy time series methods received increasing attention in many forecasting applications. To achieve better forecasting accuracy, Ref. [6] presented an effective approach which can properly adjust the lengths of intervals. Chen in [7] presented a new forecasting model based on the high-order fuzzy logical relationship groups to forecast the enrolments of the University of Alabama. Singh [9] developed a simplified and robust computational method for the forecasting rules based on one and various parameters as fuzzy logical relationships. Lee et al. in [12] presented a method for forecasting the temperature and the TAIFEX based on the high-order fuzzy logical relation groups and genetic algorithm. They also used genetic algorithm and simulated annealing in it. Recently, Particle swarm optimization technique has been successfully applied in many applications. Based on Chen's model [5] , Kuo et