Forecasting new daily confirmed cases infected by COVID-19 in Italy from April 9 ^th to May 18 ^th 2020

We aim at forecasting the outbreak of COVID-19 in Italy by using a two-part time series to model the daily relative increments. Our model is based on the data observed from 22 February to 8 April 2020 and its objective is forecasting 40 days from 9 April to 18 May 2020. All the calculations, simulations, and results in the present paper have been done in MatLab R2015b. The average curve and 80% upper and lower bounds are calculated based on 100 simulations of the fitted models. According to our model, it is expected that by May 18 ^th , 2020, the relative increment (RI) falls to the interval of 0.31% to 1.24% (average equal to 0.78%). During the last three days of the studied period, the RI belonged to the interval 2.5% to 3%. Accordingly, It is expected that the new daily confirmed cases face a decreasing to around 1900 on average. Finally, our prediction establishes that the cumulative number of confirmed cases reaches 237635 (with 80% confidence interval equal to [226340 248417] by May 18 ^th , 2020.