Implementation of Hybrid Prediction Model: An Unsupervised and Supervised Learning Perspective

Abstract

Using raw data to make inferences is the core of data science. This might be accomplished by closely examining the complex trends and patterns in the data. Machine Learning based forecasting methods have shown useful in predicting perioperative outcomes to improve the quality of future event planning decisions. In many application sectors where machine learning models were used, it has long been necessary to identify and prioritise the negative characteristics of a threat. Wish to provide precise predictions about a certain set of data, for example, use machine learning techniques in data science. Numerous prediction algorithms are now in use to address forecasting issues. Numerous epidemiological models are being employed internationally to forecast pandemic mortality rates and the number of affected people. Making the right decisions depends on the development of reliable prediction models. Epidemiological models have had trouble making longer-term forecasts with a higher degree of accuracy due to a lack of significant data and ambiguity. This research suggests a hybrid machine learning approach to anticipate the pandemic in contrast to Susceptible-Infected-Resistant based models, and we demonstrate its potential using COVID-19 data from India. Order to improve the identification of epidemics early, The model can also be updated using data from sources like search engine searches. Results from two well-known machine learning methods were compared to those from the improved SIR and SEIQR models.