"Will COVID-19 Cases in India Reach 10 Million? Forecasting Approach by Using α -Sutte Indicator Method"

COVID-19 is an infectious disease caused by a new coronavirus (SARS-CoV-2), first reported in Wuhan, Hubei Province, China in December 2019 Yang, et al. [1]. In India, COVID-19 pandemic began to spread with a first confirmed case on 30 January 2020, which was originated from China Yang, et al. [1]. Currently, India has the largest number of COVID-19 confirmed cases in Asia, after the United States, it has second highest numbers of confirmed cases in the world with 9,735,975 of confirmed people with deaths 141,398 as at 01 PM on 09 December 2020 (IST). To anticipate the many confirmed cases of COVID-19, India began lockdown with a 14-hour voluntary public curfew on 22 March 2020 which was followed by mandatory lockdowns from 24 March 2020 (Government of India, 2020). To determine the COVID-19 confirmed case rate further in the future and to know the date when India will be reported 10 million COVID-19 confirmed cases, it is important to forecast the data with an efficient prediction model. COVID-19 forecasting is necessary to determine an overview of the increase in COVID-19 confirmed cases ARTICLE INFO ABSTRACT

from time to time so that it can be taken into consideration in the decision-making process to control disease and stop transmission.
Das [4] used the SIR model to estimate the reproduction number of COVID-19 at national and state level in India. Li, et al. [5] proposed a temporal deep learning method, based on a time-aware long short-term memory (T-LSTM) neural network to predict COVID-19 for Wuhan, China. Deep Learning-based models are also used for predicting the number of novel coronavirus (COVID- 19) positive reported cases for 32 states and union territories of India Arora [6]. A logistic model and machine learning based time series prediction model were used to the prediction of COVID-19 for global, Brazil, Russia, India, Peru, and Indonesia Wang [7]. Fong, et al. [8] has proposed GROOMs which ensembles five types of forecasting methods to predict pandemic like COVID-19. Singh, et al. [9] used ARIMA model for predicting the COVID-19 disease spread trajectories for the next two months. The epidemiologic of COVID-2019 is forecasted with the ARIMA in Indian states with a high daily incidence Roy [10]. Castillo [11]  Singhal, et al. [13].
Achterberg, et al. [14] used the Network Inference-based Prediction Algorithm (NIPA) to forecast the spread of COVID-19 in Hubei, China and Netherlands and he found that network-based forecasting is superior to any other forecasting algorithm. Widely used methods like ARIMA, Holt-Winters, alpha-sutte indicator, NNAR, Theta and suttARIMA have been used in this paper for forecasting modelling. Ahmar, et al. [15] has proposed sutteARIMA method to COVID-19 in the USA. Ahmar, et al. [16] used the sutteARIMA, a short-term forecasting method to predict COVID-19 and Stock market in Spain. Saleh [17] used the SutteARIMA model to predict COVID-19 confirmed cases in Spain. Comparison of ARIMA, Holt-Winters, SARIMA, and α-Sutte Indicator has been conducted by Ahmar, et al. [18]. This study has compared α-Sutte indicator, Holt-Winters and ARIMA models and compared which one is most efficient in predicting COVID-19 confirmed cases in India. Finally, the α-Sutte indicator was found most effective method to predict COVID-19 in India, and it was used to determine whether confirmed COVID-19 cases in India would have reached 10 million.

ARIMA
The (Z t ) process is an autoregressive-moving average or ARMA (p,q) model if it fulfilled as (Wei,1994): If there is a differencing then the ARIMA model becomes as follows:

α-Sutte Indicator
The α-Sutte indicator is based on the practice of forecasting that is developed on the previous values of the variable or data set Ahmar, et al. [19].

Holt-Winters
Holt-Winters's prediction method has been categorized into two part: Multiplicative Holt-Winters (MHW), and Additive Holt-Winters (AHW). The equation of MHW is described in the following The component form for the additive method is: Where: yt = data on t time, s= the seasonal length in a certain time, and m=the amount of data to be predicated.

Results and Discussion
( Figure 1) shows that total COVID-19 confirmed cases have been an upward trend from the beginning. This trend indicates that 10 million cases will be completed in a short time period. This is also reinforced by (Figure 2), which explains that the average number of confirmed cases per day in India is around 35000 (in last five days). To further study the likelihood of succeeding the 10 million confirmed cases, it is necessary to forecast the data. In this study, the α-Sutte indicator method had an accuracy rate on data fitting of 0.025% for the period of 30 November 2020 to 06 December 2020. (Table 1) Finally, the α-Sutte indicator method is used to predict data ( Table 2). The α-Sutte indicator predicts that 10 million cases will be reported on 15 or 16 December 2020 in India situation report 42 by the WHO (Figure 3).