Hybrid Mathematical Model for Data Classification and Prediction: Case study COVID-19

The increase of data availability has stimulated researchers to benefit from this data in predicting the hidden pattern for knowledge discovery. Data classification and machine learning algorithms are becoming important tools used in knowledge discovery. In this paper, we propose a hybrid classification model that combines some features and parameters from a probabilistic model and some other parameters from a divide and conquer model in a linear one. In our model, we generate a set of functions related to the number of attributes and the value of each attribute. Afterwards, these functions are reduced according to the number of classes needed. We test our model on collected data about symptoms in people infected with COVID-19 in England. Our simulation results show an accuracy rate in the range 50-80%. We expect to increase the accuracy rate if we increase the size of data used or we increase the number of attributes.