Journal of Xidian University https://doi.org/10.37896/jxu14.

4/164 ISSN No:1001-2400

A Markov Model for Prediction of Corona Virus

COVID-19 in India- A Statistical Study
Dr.R.Arumugam#1, M.Rajathi*2
#1
Department of Mathematics
*2
Department of Education
Periyar Maniammai Institute of Science & Technology

Abstract— A Markov chain model is mainly used for business, manpower planning, share market and many different areas. Because
the prediction of the any ideas based on the Markov chain the result need to be efficient. Now, the infection of corona virus COVID-
19 is a large task for the human being as well as the government. This paper is focusing tool for prediction of corona
virus infection with a markov chain model. Markov chain model had been used to predict the corona virus (COVID-10) based at
the secondary data as on 13th March 2020. The 1st order markov models had been used to predict the impact of corona virus
using probability matrices and Monte Carlo simulation. To present the applications of this model, 2020 corona virus pandemic in
India by country and union territory become used as a case study. It will be useful for prediction of the corona virus COVID-19
in destiny.
Keywords— Markov chain, transition probability, COVID-19, forecasting, Monte Carlo simulation

1. INTRODUCTION

In the modern world, there are so many methods to solve the complex problem using forecasting
approach. Some of the forecasting strategies are Seasonal Autoregressive Integrated Moving Average
(SARIMA), Autoregressive Moving Average Model (ARIMA), and Artificial Neural Networks (ANN)
etc., Markov chain is an essential tool that has been prepared to resolve complex problem such as peak
power utilization [1]. The special case of the stochastic model for the complex problem is Markov chain
model [2]. Nowadays to assess the transition matrix from the given state system were used numerous
techniques [3]. This is a stochastic processes it is contain all the details about the future. Also the vital
components in developing the markov model are transition state and transition matrix; both of which will
sum up all the important parameter of active stage. Moreover the markov chain, particularly the behavior
of the direction. This kind of model is based on the maximum likelihood method and the system of linear
programming and additionally the studies that used the markov chain model as a trend (ie forecasting tool)
for the API (Air Pollution Index) are scant. Therefore the model is depends upon on the linear
programming and maximum likelihood technique. There are so many studies are there for the air pollution
the use of prediction [4-8]. In the particular area to assess the amount of water conventional a place to get
collectively an industry, agriculture and different activities related to humans [9]. The rainfall confirmation
analysis for extensive duration gives the data approximately the rainfall stages and variability [10]. The
prevalence of the probability in the prediction of rainfall has stayed abandoned in spite of its huge
relevance in hydrologic threat and dependability studies.
The deviation of rainfall is very crucial for agriculture. The evaluation of yearly rainfall and its
prevalence from the existing time series to get the perdition for the once a year rainfall using statistical
parameters like mean, standard deviation and the coefficient of variation. Numerous techniques [9-13]
have been proposed through various researchers for rainfall modeling information. A markov model for
prediction of annual rainfall have been proposed by A.O.Ibeje et.al.,[14]. Applications of Manpower
Levels for Business with Various Recruitment Rates in an Eleven point State space through Stochastic
Models have been proposed by R.Arumugam et. al., [15].

2. METHOD AND MATERIALS

2.1 STUDY AREA

The prediction of the impact of corona virus COVID-19 in India from February to March 2020 had
been calculated all the states the using Markov chain stochastic model. The stochastic model is familiar to
discover the trend values based on the markov chain.

VOLUME 14, ISSUE 4, 2020 1422 http://xadzkjdx.cn/

Journal of Xidian University https://doi.org/10.37896/jxu14.4/164 ISSN No:1001-2400

TABLE 1
CORONA VIRUS PANDEMIC IN INDIA BY STATE AND UNION TERRITORY AS ON 30 TH MARCH 2020

State or Union Total

S.No territory Active cases Recoveries Deaths cases
1 Maharashtra † 215 25 7 247
2 Kerala † 202 20 1 223
3 Karnataka 83 5 3 91
4 Delhi † 72 6 2 80
5 Uttar Pradesh † 72 11 0 83
6 Telangana † 70 1 1 72
7 Gujarat 69 1 6 76
8 Rajasthan † 60 3 0 63
9 Tamil Nadu † 67 4 1 72
10 Madhya Pradesh 39 0 2 41
Jammu and
11 41 1 1 43
Kashmir
12 Punjab 39 1 1 41
13 Haryana † 35 17 0 52
14 West Bengal 22 1 0 23
15 Andhra Pradesh 23 1 0 24
16 Bihar 15 0 1 16
17 Ladakh 13 3 0 16
Andaman and
18 10 0 0 10
Nicobar island
19 Chandigarh 9 0 0 9
20 Chhattisgarh 7 0 0 7
21 Uttarakhand 7 2 0 9
22 Goa 5 0 0 5
23 Himachal Pradesh 3 1 1 5
24 Odisha 3 0 0 3
25 Manipur 1 0 0 1
26 Mizoram 1 0 0 1
27 Puducherry 1 0 0 1
Total 1184 103 27 1314

2.2 DATA COLLECTION

The secondary data were collected from the website, especially https://www.mygov.in/covid-19
website as on 30th March 2020 for the study on “A MARKOV MODEL FOR PREDICTION OF CORONA VIRUS COVID-19 IN
INDIA”.
We followed statistical tools like descriptive statistics, viz., mean, median, mode, range, standard
deviation, skewness and kurtosis for the Corona pandemic COVID-19.

Fig 1: Impact of COVID-19 on Indian Economy Fig 2: Corona virus image

VOLUME 14, ISSUE 4, 2020 1423 http://xadzkjdx.cn/

Journal of Xidian University https://doi.org/10.37896/jxu14.4/164 ISSN No:1001-2400

Mean, median, mode, range, standard deviation and variance are focused on the general public
residing in India at some stage in the pandemic of COVID- 19 and the skewness and
kurtosis are spotlighted the dearth of symmetry to represents the extent of COVID- 19.

Fig 3: Symptoms of Corona virus Fig 4: How it spreads Corona virus

Fig 5: Prevention of Corona virus

2.3 FORMULATION OF THE MODEL

Let X 0 , X 1 , X 2 ,......, X n be a random variables with times t 0 , t1 ,......,t n is stated to be a markov
method and it's satisfy the following property:
  
P X p 1  X p1 / X p  X 0 , X p1  X 1 ,........X p  X n = P X p 1  X p 1 / X p  X n  (1) 
The above belongings is also referred to as one-step transition possibility from one state i at t-1 to t. i.e.,
Pij . . By the definition of probability,
n
0  Pij  1; where, both i, j = 1, 2, ….., n and P
j1
ij  1          (2)

Pij  Pij   Pij ; i, j  1,2,3,.........n          (3)

Where, Pij the no. of instances is the observed information from nation i to j.
The first order markov chain of the probability transition matrix P is,
 p11 p12 p13 ...... p1n 
p p 23 
 21 p 22 ...... p 2 n 
P  . . . ...... .           (4)
 
. . . . . 
 p n1 pn2 pn3 ...... p nn 


3. ANALYSIS

For generating the sequences of corona virus (COVID-19) states, the beginning state we are saying that
i selected arbitrarily. The random values between 0 and 1 were produced by way of the use of random
number generator. This is also referred to as Monte Carlo simulation. The cumulative possibility transition
matrix was received by using successive multiplication of P matrix with the aid of itself till a stabilization
of the transition probabilities led to the transition probability matrix. Thus if the transition possibility
inside the ith row on the jth state was Pij . , then the cumulative probability is

VOLUME 14, ISSUE 4, 2020 1424 http://xadzkjdx.cn/

Journal of Xidian University https://doi.org/10.37896/jxu14.4/164 ISSN No:1001-2400

i
Pij   Pij            (5)
j 1
TABLE 2
DESCRIPTIVE STATISTICS OF CORONA VIRUS INFECTION IN INDIA AS ON 30 TH MARCH 2020
Bootstrapa
95% Confidence
Interval
Statistic Std. Error Bias Std. Error Lower
Active_cases N 27 0 0 27
Range 214
Minimum 1
Maximum 215
Sum 1184
Mean 43.85 10.527 -.59 10.22 25.52
Std. Deviation 54.700 -3.687 12.956 25.383
Variance 2992.131 -222.052 1304.744 644.301
Skewness 2.141 .448 -.238 .670 .295
Kurtosis 4.820 .872 -.447 3.587 -1.608
Recoveries N 27 0 0 27
Range 25
Minimum 0
Maximum 25
Sum 103
Mean 3.81 1.279 -.08 1.20 1.67
Std. Deviation 6.645 -.410 1.516 2.645
Variance 44.157 -2.987 18.251 6.997
Skewness 2.223 .448 .037 .674 1.106
Kurtosis 4.214 .872 .796 4.215 -.348
Deaths N 27 0 0 27
Range 7
Minimum 0
Maximum 7
Sum 27
Mean 1.00 .342 -.01 .35 .37
Std. Deviation 1.776 -.136 .502 .636
Variance 3.154 -.213 1.598 .405
Skewness 2.492 .448 -.222 .739 .880
Kurtosis 6.140 .872 -.452 4.618 -.454
Valid N (listwise) N 27 0 0 27

TABLE 3
FREQUENCY DISTRIBUTION OF CORONA VIRUS INFECTION AS ON 30TH MARCH 2020
State Status Frequency
1 Active cases 1184
2 Recoveries 103
3 Deaths 27
Total 1314

The probability transition matrix P describes the Markov chain representing three states of corona virus
(COVID-19) infection was obtained from (4) as
0.9331 0.0483 0.0186 
P  0.6628 0.2791 0.0581          (6)
0.4681 0.1064 0.4255 
The cumulative probability transition matrix Pcu attained from the equation (3) as,
0.9331 0.9814 1.0
P  0.6628 0..9419 1.0            (7)
0.4681 0.5745 1.0
Monte Carlo simulation states are,

VOLUME 14, ISSUE 4, 2020 1425 http://xadzkjdx.cn/

Journal of Xidian University https://doi.org/10.37896/jxu14.4/164 ISSN No:1001-2400

 A if v  0.9331

State =  B if 0.4681  v  0.6628          (8)
C if v  0.4681

4. DISCUSSION AND RESULTS

Table 1 represents that the Corona virus pandemic in India by means of state and union territory as
on thirtieth March 2020 with three levels like active cases, recoveries and deaths in Indian states. Figure 1
indicates that the Impact of COVID-19 on Indian Economy and the second figure offers that the images of
Corona virus. Third figure demonstrate that the symptoms of Corona virus and fourth figure gives an idea
about how spreads Corona virus and figure five illustrates that the prevention of Corona virus.
Table 2 illustrates that the infection of Corona Virus (COVID-19) as on 30th March 2020 in India.
The range of corona between 1 and 215 within the active stage with mean 43.85 and the standard deviation
is 54.7. The skewness represents that the dearth of symmetry (ie., 2.141) and the kurtosis is more than 3.
Therefore the skewness is positive, it is shows that lepto kurtic, this means that that the probability plot has
peaked at the right side. Similarly the recoveries and death stage represents that positive skewness and
leptokurtic with N= 27.
The transition chances illustrates that, there is 93% chance in the first level (Active stage) of
corona if N=1314. In the second stage (Recoveries) 66% chances are there. Therefore Corona recovered
possibilities in India as on 30th March is 66%. In the third stage (Death level) 46.8% of probabilities of
there for the given statistics if N=1314.
From equation (8), Monte Carlo simulation, the alternate of the corona virus (COVID-19) in which
generated as the random range adjustments values and assumes the distinct states of corona infection;
where v is a random wide variety of uniform distribution.
5. CONCLUSION

This study results revealed that markov chain is beneficial in simulating the corona infection in
numerous stages. This type of simulation could be very much useful in generating the time period of
corona virus infection. The evaluation of corona infection indicates that markov chain approach offers us
one opportunity of modeling in future.
REFERENCES

[1] Elgharbi, S.; Esghir, M.; Ibrihich, O.; Abarda, A.; El Hajji, S.; Elbernoussi, S. Grey-Markov Model for the Prediction of the Electricity Production and
Consumption. In Lecture Notes in Networks and Systems; Springer New York: New York, NY, USA, 2020; Volume 81, pp. 206–219.
[2] Winston, W.L.; Goldberg, J.B. Operations Research: Applications and Algorithms; Thomson Brooks/Cole: Belmont, CA, USA, 2004.
[3] Chatfield, C. Statistical Inference Regarding Markov Chain Models. J. R. Stat. Society. Ser. C (Appl. Stat.) 1973, 22, 7–20. [CrossRef]
[4] Ibarra-Berastegi, G.; Elias, A.; Barona, A.; Saenz, J.; Ezcurra, A.; Diaz de Argandoña, J. From prognosis to analysis for forecasting air pollutants the
usage of neural networks: Air pollutants monitoring in Bilbao. Environ. Model. Softw. 2008, 23, 622–637. [CrossRef]
[5] Yan Chan, K.; Jian, L. Identification of substantial factors for air pollution levels using a neural network based understanding discovery system.
Neurocomputing 2013, 99, 564–569. [CrossRef].
[6] Fernando, H.J.S.; Mammarella, M.C.; Grandoni, G.; Fedele, P.; Di Marco, R.; Dimitrova, R.; Hyde, P. Forecasting PM10 in metropolitan areas: Efficacy
of neural networks. Environ. Pollut. 2012, 163, 62–67. [CrossRef] [PubMed]
[7] Güler, N.; Güneri ˙I¸sçi, Ö. The local prediction version of PM10 concentrations for Turkey. Atmos. Res. 2016, 180, 64–77.
[8] Ong, B.T.; Sugiura, K.; Zettsu, K. Dynamically pre-skilled deep recurrent neural networks using environmental tracking statistics for predicting PM2.5.
Neural Comput. Appl. 2016, 27, 1553–1566. [CrossRef] [PubMed]
[9] C.W. Richardson (1981). Stochastic simulation of each day precipitation, temperature, and sun radiation, Water Resour. Res. Vol. 17, No. 1, pp.182-190.
[10] A.O. Nnaji. Forecasting seasonal rainfalls for agricultural decision making in northern Nigeria. J. Of Agric. And Forest Met. Vol. 107, 2001, pp.1. 193-
205.
[11] R. Lazaro, F.S. Rodrigo, L. Gutierrez, J.Puigdefafregas. Analysis of a 30-12 months rainfall record (1967-1997) in Semi-arid SE Spain for implications
on vegetation, J. Arid Environ.. Vol.48, 2001, pp. 373 - 395.
[12] S. Tamil Selvi and R. Samuel Selvaraj. Stochastic Modelling of Annual Rainfall at Tamil Nadu. Universal Journal of Environmental Research and
Technology, Volume 1, No. 4, 2011, pp. 566-570.
[13] M.Thyer and G. Kuzcera. Modelling long-term patience in rainfall time series: Sydney rainfall case study. Hydrology and Water Resources Symposium,
Institution of Engineer, Australia, 1999, pp.550-555.
[14] A.O. Ibeje, J.C.Osuagwu and O.R. Onosakponome. A Markov Model for Prediction of Annual Rainfall, International Journal of Scientific Engineering
and Applied Science (IJSEAS), Vol.3, Issue-11 Nov 2018, pp 1-5.
[15] R.Arumugam, M. Rajathi, Applications of Manpower Levels for Business with Various Recruitment Rates in an Eleven point State space through
Stochastic Models, Compliance Engineering, Vol. 10, Issue 12, Dec 2019, pp. 384-392.

VOLUME 14, ISSUE 4, 2020 1426 http://xadzkjdx.cn/