Materials Today: Proceedings xxx (xxxx) xxx

Contents lists available at ScienceDirect

Materials Today: Proceedings

journal homepage: www.elsevier.com/locate/matpr

A simple mathematical model to predict and validate the spread of

Covid-19 in India
Harish Kumar a,⇑, Pawan K. Arora b, Meena Pant a, Anil Kumar c, Shahroz Akhtar Khan b
a
National Institute of Technology Delhi, India
b
Galgotias College of Engineering and Technology, Gautambudh Nagar, U.P, India
c
H.R.Group of Institutions, Ghaziabad, U.P, India

a r t i c l e i n f o a b s t r a c t

Article history: The new outbreak of the corona virus (Covid-19) is expanding rapidly worldwide, disrupting millions and
Received 26 October 2020 prompting authorities to take swift measures to avoid the disease. National lockdown imposed by the
Received in revised form 5 March 2021 Indian government since 25 March 2020, the early lockdown action shows as compared to many other
Accepted 12 March 2021
Countries/states can benefit from limiting the final size of the epidemic. A report on the issue of spreading
Available online xxxx
the Covid-19 modeling in India is under review. This study analyzes Covid-19 infections by 20Dec 2021
and presents a mathematical approach for forecasting new cases or cumulative cases in practical situa-
Keywords:
tions. This forecast is much needed to schedule/continue medical set-ups for possible action to tackle the
Pandemic
Covid-19
Covid-19 outbreak. It is important to mention here that the number of authors has proposed different
Prediction models for predicting the expansion of Covid-19 to India and other countries; almost no model has yet
Infection to be demonstrated viable. With this mathematical model, it is simple to forecast the transfer of
Mortality Covid-19. It is clear from the data that lockdown has played a significant role in controlling the transmis-
Modeling sion of the disease. A close match between the predicted empirical results and the available results
proves the derived model similarity.
Ó 2021 Elsevier Ltd. All rights reserved.
Selection and peer-review under responsibility of the scientific committee of the 3rd International Con-
ference on Computational and Experimental Methods in Mechanical Engineering.

1. Introduction about 10 times fatal [9]. Fig. 1 illustrates the countries reporting
Covid-19 cases in the globe. It can see that Covid-19 cases are reg-
As reports showed, a pandemic of unknown pneumonia started istered in each country except in two nations around the world
in Wuhan, China, in December 2019, but there are reports that it [10]. Fig. 2 shows a map of India and various states of India, with
began by the end of November 2019 [1,2]. A new variant of the cor- an estimated number of Covid-19 infections reported to date [11].
ona virus COVID-19 similar to SARS-CoV, for which health experts
have approved no valid medicine or vaccine. This virus is transmit-
1.1. Present scenario
ted by the respiratory tract or by contact with infected particles or
pathogenic organisms over an incubation period of 2 to 14 days
It is essential to determine the needed potential facilities, which
[3,4]. Until then, the world could be protected by various preven-
may be necessary for terms of hospitals, beds, ventilators, medici-
tive measures; COVID-19 pandemic has taken over the entire
nes, etc. In regards to health personnel including doctors, nurses,
globe. Over an early duration of 4 months, many as five million
and support personnel, a variety of researchers, including physi-
cases are publicly revealed, with much more than 300,000 causal-
cists, biologists work for thorough evaluation and further conse-
ities in over 150 countries [5,6]. COVID-19 is not explicitly treated,
quences of their developed model in the study of infected data,
and the mortality rate varies across different countries and ranges
recovered data, and death rates. Infection, recovery, and mortality
from 2 to 15 percent [7,8]. The epidemiological data verify that the
rates depend not only on the form of Covid-19 but also on the
spread rate is very high, and the WHO confirmed that SARS-CoV is
country’s reaction. In this scenario, the researcher would like to
examine data on infected, recovered, and dead people in India
⇑ Corresponding author. directly compared with other countries and the worldwide mean.
E-mail address: [email protected] (H. Kumar). The specific modeling was designed and evaluated by a few

https://doi.org/10.1016/j.matpr.2021.03.434
2214-7853/Ó 2021 Elsevier Ltd. All rights reserved.
Selection and peer-review under responsibility of the scientific committee of the 3rd International Conference on Computational and Experimental Methods in Mechanical
Engineering.

Please cite this article as: H. Kumar, P.K. Arora, M. Pant et al., A simple mathematical model to predict and validate the spread of Covid-19 in India, Materi-
als Today: Proceedings, https://doi.org/10.1016/j.matpr.2021.03.434
H. Kumar, P.K. Arora, M. Pant et al. Materials Today: Proceedings xxx (xxxx) xxx

Fig. 1. Countries reported Covid-19 infections so far [10].

Rashad Eletrey (2020) investigates the impact of evolutionary

adaptations, referring COVID-19 on the transmission of processes
in complex networks [19].
GianlucaMalato (2020) evaluated the pattern of infected per-
sons using a mathematical model based on a logistic model (equa-
tion 1) and an exponential model (equation 2) for publicly
available data through the portal of the Government of Italy. The
data was compared to the actual data available for far, and a critical
consistency is obtained with significant deviation. X is the time in
the equation, a, b, c are parameters/variables [20].
Tain-Mu-Chen, et. Al. (2020) provided a mathematical model is
used to simulate the transmissibility process of the novel coron-
avirus in China. He used Berkeley Madonna 8.3.18 software (devel-
oped by R Macey and G Oster at Berkeley University of California)
to match the data collected from the competent Chinese authority,
which is in the public domain [21].
In a recent work researchers have analyzed the global impact of
Covid-19 with emphasis given on daily cumulative infections and
cumulative deaths. The data for Europe and Americawere analyzed
and impact was judged in relation to the data for the worldfor
Covid-19 outbreak [22].
However, due to limitations, only specific mathematical models
are listed here, but at the same time, researchers and practitioners
are seeking to create/examine current statistical/mathematical
methods in times of hardship for the human race. On further obser-
Fig. 2. States of India showing Covid-19 infections so far [11].
vations, authors have found that the models proposed by various
researchers are focused on the constraints depending on a specific
city o country and are not accurate for each area. Different con-
stants assumed/evaluated in different models, as discussed above,
researchers, and their findings were presented [12–15]. Some of are very specific, limiting the use of models as general-purpose
the models are discussed below, along with their brief concept models. Addressing the limitations, authors present simple mathe-
and shortcomings. matical models based on extensive data available in India (the first
Julia R Gog (2020) explained the classic epidemic susceptible-i case of Covid-19 reported in India on 30 January 2020, the con-
nfectious-recovered (SIR) model, a very simple model proposed firmed cases then reached to 100,000 on 19 May 2020 and-
in the 1920 s. It is called the SIR model, and S is the number of peo- crossed11046000 cases up to now), which do not include
ple who are susceptible to disease but have not been infected so different particular constants / computed factors [11]. The relevant
far, I form the number of people who are ill, and R are the people data and subsequent evaluation of the mathematical model are
who are either dead or healed from infection [16,17]. SanglierCon- described in the following section.
trearas (2020) suggested the prediction of COVID-190 s epidemio-
logical progress through modeling, analysis, and interpretation of
data, and recommending various quadratic equations and appears 2. Methodology
to fit for critical data evaluation. The multiple factors studied and
evaluated are the number of people infected, the death rate, the Fig. 3 gives a brief glimpse of Covid-19 spread in India. One can
number of people recovered, the duration of treatment, etc. [18]. find the confirmed cases of Covid-19 in India as on 20 Dec 2020.
2
H. Kumar, P.K. Arora, M. Pant et al. Materials Today: Proceedings xxx (xxxx) xxx

Fig. 3. Impact of Covid-19 in India since reporting of first case until Dec20, 2020.

The data for each state has been shown in tabulated form in table Authors have attempted a number of different generalised
1. The same information of cumulative sum of different parameters mathematical models to predict some function for cumulative
since inception for India are shown in Fig. 3. Out of the total infections, cumulative recoveries, cumulative deaths and cumula-
infected people so far, the cured, deaths and infected are repre- tive active infections, few of them areQuartic / bi-quadratic regres-
sented in form of pie-chart through Fig. 4. sion,exponential function, sum of sine function, Gaussian model,

Table 1
Statistics of the Covid-19 information as on Dec20, 2020 [11].

S. No. State / Union Territory Total cases Deaths Recoveries Active cases
1 Andaman and Nicobar Islands 4881 61 4732 88
2 Andhra Pradesh 878,723 7076 867,445 4202
3 Arunachal Pradesh 16,630 55 16,350 225
4 Assam 215,409 1017 210,869 3523
5 Bihar 247,244 1352 240,915 4977
6 Chandigarh 19,073 308 18,328 437
7 Chhattisgarh 267,219 3181 247,480 16,558
8 Dadra and Nagar Haveli and Daman and Diu 1640 1 1632 1
9 Delhi 617,005 10,277 596,580 10,148
10 Goa 50,064 721 48,371 972
11 Gujarat 235,299 4234 219,225 11,840
12 Haryana 257,644 2821 248,935 5888
13 Himachal Pradesh 52,329 864 46,221 5196
14 Jammu and Kashmir 118,263 1841 112,568 3854
15 Jharkhand 113,025 1010 110,307 1708
16 Karnataka 909,469 12,009 882,944 14,497
17 Kerala 705,870 2817 641,285 61,600
18 Ladakh 9279 124 8800 355
19 Lakshadweep 0 0 0 0
20 Madhya Pradesh 231,284 3481 216,485 11,318
21 Maharashtra 1,896,518 48,746 1,783,905 62,743
22 Manipur 27,648 337 25,646 1665
23 Meghalaya 13,253 134 12,620 499
24 Mizoram 4122 7 3963 152
25 Nagaland 18 0 9 0
26 Odisha 8901 29 8765 107
27 Puducherry 37,748 626 36,777 345
28 Punjab 163,042 5201 152,223 5618
29 Rajasthan 298,996 2617 283,957 12,422
30 Sikkim 5573 124 4981 374
31 Tamil Nadu 806,891 11,983 785,315 9593
32 Telangana 281,414 1513 273,013 6888
33 Tripura 33,164 375 32,509 257
34 Uttarakhand 2961 28 2443 490
35 Uttar Pradesh 2170 20 2068 82
36 West Bengal 536,828 9360 509,697 17,771
Total 9,069,597 134,350 8,657,363 276,393

3
H. Kumar, P.K. Arora, M. Pant et al. Materials Today: Proceedings xxx (xxxx) xxx

Fig. 4. Distribution of Covid-19 infections in India as on Dec 20, 2020.

Fourier series and sum of exponential function series with confi- been madeto 20 days forward on the basis of these model equa-
dence level of 95%. But, there was not a single polynomialorder tions. The plot of the same has been shown in Fig. 5 while Fig. 6
or function, which could be helpful in present scenario. Authors is supplied to compare the results of model equations with the
are hereby proposing the use of generalized polynomialregression actual plotted data for Covid-19 infections for first 115 days.
of order 6 and discussing the function equations for cumulative If we carefully analyse and compare the Fig. 5 with Fig. 3, we
infections, cumulative recoveries, cumulative deaths and cumula- will see that on Day 130, (Aug13, 2020), the line showing number
tive active infections respectively as a function of time.Data for dif- of cumulative active infected people is changing slope in having a
ferent functions as mentioned earlier are considered for first small peak, thereafter the number will keep of reducing which
115 days of spread of Covid-19 in India, i.e., during April 01 – indicates that the recovering people are more than the freshly
Jul24, 2020. Table 2 summarizes the equation for each function infected people. It has been found that the projected and the actual
with the R-squared value for better understanding. The model data are in the good agreement exhibiting the similar pattern.
equations (as a function of time in terms of days) are used to plot
the existing data from April 01 – Jul24, 2020and projection has 3. Results and discussions
Table 2
Equation for different functions.
This paper has attempted to investigate the spread of Covid-19
in India on the basis of information available through authenti-
S. Function for Model Equation R- cated sources. India reported its first Covid-19 infection on January
No. Squared
30, 2020 and for about five weeks, there were very little or negligi-
1. Cumulative 1E-06 * x6  0.0007 * x5 + 0.1751 * x4– 0.9998 ble infections reported across the country. During this period, lot of
infections (CI) 17.607 * x3 + 791.72 * x2–10431 * x
2. Cumulative 8E-07 * x6  0.0006 * x5 + 0.1726 * x4– 0.9999
careful considerations were made by Government of India like
recoveries(CR) 18.97 * x3 + 912.71 * x2–14113 * x thermal screening at airports, awareness among the citizens
3. Cumulative 9E-09 * x6  6E-06 * x5 + 0.0015 * x4  0.9998 regarding Covid-19 and its spread, sanitization of probable or con-
deaths (CD) 0.1376 * x3 + 7.4558 * x2–97.284 * x firmed zones of Covid-19 infection etc. The Covid-19 infections
4. Cumulative 1E-07 * x6  5E-05 * x5 + 0.0011 * 0.9795
started surging (per day reporting of fresh infections in two digits)
active infections x4 + 1.4914 * x3  128 * x2 + 3772.5 * x
(AI) in March 2020 and the Government of India imposed nation-wide
lock-down with minimal / emergent connectivity to the rest of

Fig. 5. Projection of Function on the basis of model equations Vs span of time (days)upto Aug 13, 2020 (day 135).

4
H. Kumar, P.K. Arora, M. Pant et al. Materials Today: Proceedings xxx (xxxx) xxx

Fig. 6. Actual data of Covid-19 infectionVs span of time (days)upto Jul 24, 2020 (day 115).

world and within country, there was no transport permitted with (b) Brief discussion regarding the recent work reported by vari-
complete stop to industry, institutions, retails etc. excluding ous researchers through the globe is presented along with
health, law and order sector. the mathematical models developed / implied by them.
Initially, as evident from the international databases, there (c) A simple mathematical model for assessment of Covid-19
Covid-19 infections grew in small number, but because of its prop- infections in India has been presented using generalized
agating nature, soon the infections keep on rising and the country polynomial regression. The model equations for Cumulative
extended the lock-down on April 15, May 04 and May 18, 2020 infections (CI), Cumulative recoveries (CR), Cumulative
respectively. As of now, about 972,121,246 Covid-19 infections deaths (CD), and Cumulative active infections (AI) are
reported so far (Day 264) with a propagation rate of approximately derived and the projected values are compared with infor-
5% per day. The databases have been used to investigate the pat- mation available regarding the same.
tern of spread of Covid-19 infections and the following parameters (d) The Covid-19 infections are seen surging in the country from
are investigated: day 130 and a local peak of infections is expectedwith pro-
jections of decay in infection rate thereafter. The same can
(a) Cumulative infections, CI (derived from infections reported be validated through actual data plotted in Fig. 4 near to
each day) the day 130.
(b) Cumulative recoveries, CR (derived from recoveries reported (e) The predicted data for CD is found in close agreement with
each day). the actual CD value and its pattern. Predicted CD pattern
(c) Cumulative deaths, CD (derived from deaths reported each has a little rise in slopebeyond the day 130,showing increase
day) in cumulative deaths in the country per day. The similar pat-
(d) Cumulative active infections, AI (derived from active CI, CR tern can be seen in the actual data for CD plotted in Fig. 4.
and CD respectively)
CRediT authorship contribution statement
Authors attempted to investigate different models reported and
applied across the globe for prediction of spread of Covid-19 and Harish Kumar: Investigation. Pawan K. Arora: Methodology.
felt that the models like SIR, SIRD and SIER, though have been Meena Pant: Writing - original draft. Anil Kumar: Writing -
applied through the globe, but they are dependent over various review & editing. Shahroz Akhtar Khan: Writing - original draft.
factors depending upon the circumstances over that location and
may not be applied unilaterally to any situation. As of now, there
Declaration of Competing Interest
has been enormous information available regarding Covid-19
infection, its spread; it is now possible to employ simple mathe-
The authors declare that they have no known competing finan-
matical models to predict the same. Even in India, such models
cial interests or personal relationships that could have appeared
had been used by various researchers, but either their projections
to influence the work reported in this paper.
were very much conservative or too extrapolated, and hence did
not able to present a practically viable model.
Authors have implied generalized regression model for critical References
investigation of Covid-19 infection information available regarding
[1] J. Cui, F. Li, Z.-L. Shi, Origin and evolution of pathogenic coronaviruses, Nat. Rev.
India and has used the information of 115 days to predict the Microbiol. 17 (3) (2019) 181–192.
future spread of Covid-19.Fig. 5 can be employed to make esti- [2] Muhammad Adnan Shereen, Suliman Khan, Nadia Bashir and Rabeea Siddique,
mates/projections for Covid-19 infections on days in advance. COVID-19 infection: Origin, transmission, and characteristics of human
coronaviruses, J. Adv. Res. 24 (7) (2020) 91–98.
[3] N. Zhong, B. Zheng, Y. Li, L. Poon, Z. Xie, K. Chan, et al., Epidemiology and cause
4. Conclusions of severe acute respiratory syndrome (SARS) in Guangdong, People’s Republic
of China, in February, 2003, Lancet 362 (9393) (2003) 1353–1358.
[4] N. Wang, X. Shi, L. Jiang, S. Zhang, D. Wang, P. Tong, et al., Structure of MERS-
The following conclusions have been drawn from the investiga- CoV spike receptor-binding domain complexed with human receptor DPP4,
tion presented herewith: Cell Res. 23 (8) (2013) 986.
[5] https://www.worldometers.info/coronavirus/ (assessed on May 20, 2020).
[6] https://ourworldindata.org/mortality-risk-covid (assessed on May 20, 2020).
(a) A detailed assessment of Covid-19 infection and related [7] https://www.healthline.com/health/coronavirus-covid-19 (assessed on May
aspects regarding India has been presented. 20, 2020).

5
H. Kumar, P.K. Arora, M. Pant et al. Materials Today: Proceedings xxx (xxxx) xxx

[8] Ivan Fan-Ngai Hung, Kwok-Cheung Lung, Eugene Yuk-Keung Tso et.al., [15] Jaspreet Singh, P. K. Ahluwalia and Ashok Kumar, Mathematical Model Based
Triple combination of interferon beta-1b, lopinavir–ritonavir, and ribavirin COVID-19 Prediction in India and its Different States, 10.1101/
in the treatment of patients admitted to hospital with COVID-19: an open- 2020.05.16.20104232
label, randomised, phase 2 trial, The Lancet, 10.1016/ S0140-6736(20) [16] J.R. Gog, How you can help with COVID-19 modelling, Nat Rev Phys (2020),
31101-6. https://doi.org/10.1038/s42254-020-0175-7.
[9] http://www.ucihealth.org/blog/2020/04/why-is-covid19-so-dangerous [17] https://www.maa.org/press/periodicals/loci/joma/the-sir-model-for-spread-
(assessed on May 20, 2020). of-disease-the-differential-equation-model (assessed on May 20, 2020).
[10] https://www.cdc.gov/coronavirus/2019-ncov/global-covid-19/world-map. [18] G. Sanglier Contreras, M. Robas Mora, P. Jimenez Gómez, Estimation of the
html (assessed on May 20, 2020) Epidemiological Evolution Througha Modelling Analysis of the Covid-19,
[11] https://en.wikipedia.org/wiki/COVID-19_pandemic_in_India (assessed on May Outbreak, Acta Sci. Mic. 3 (4) (2020) 152–158.
20, 2020). [19] Rashad Eletreby, Yong Zhuang, Kathleen M Carley, Osman Yağan, and H
[12] Kaustuv Chatterjee, Kaushik Chatterjee, Arun Kumar and Subramanian Vincent Poor, The effects of evolutionary adaptations on spreading processes
Shankar, Healthcare impact of COVID-19 epidemic in India: A stochastic in complex networks, PNAS 117 (1) (2020) 5664–5670.
mathematical model, Med J Armed Forces India, 10.1016/j.mjafi.2020.03.022. [20] https://towardsdatascience.com/covid-19-infection-in-italy-mathematical-
[13] Tarun Bhatnagar, Sandip Mandal, Nimalan Arinaminpathy, et al., Prudent models-and-predictions-7784b4d7dd8d (retrieved on April 14, 2020)
public health intervention strategies to control the coronavirus disease 2019 [21] T. Chen, J. Rui, Q. Wang, et al., A mathematical model for simulating the phase-
transmission in India: A mathematical model-based approach, Indian J. Med. based transmissibility of a novel coronavirus, Infect Dis Poverty 9 (2020) 24.
Res. 151 (2) (2020) 190–199. [22] Ayush Kumar Agrawal, P.K. Arora, MusarratNafees, Shahroz Akhtar Khan,
[14] Jyoti Bhola, Vandana Revathi Venkateswaran and Monika Koul, Corona Harish Kumar,Assessment of health infrastructure in tackling COVID-19: A
Epidemic in Indian context: Predictive Mathematical Modelling, 10.1101/ case study of European and American scenario,Materials Today:
2020.04.03.20047175. Proceedings,2021,ISSN 2214-7853,10.1016/j.matpr.2021.01.916.