A Project Report on Omni-platform for Covid-19: Integrated Forecasting, Symptom Analysis, and Yogasana for Mental Health Support

The COVID19 pandemic has exposed critical gaps in global healthcare systems while simultaneously affecting physical and mental well-being worldwide. While numerous digital platforms emerged to track viral spread, few addressed the comprehensive needs of individuals including predictive analytics, self-assessment capabilities, and integrated wellness interventions. This study presents a comprehensive omni-platform that integrates three complementary modules: (i) COVID19 time-series forecasting utilizing ARIMA, Prophet, and RNNLSTM models with enhanced comparative analysis; (ii) a machine learning-powered COVID symptom checker employing Bayesian inference for risk assessment; and (iii) a computer vision based Yogasana recognition system designed to support stress reduction and immune enhancement. The system was evaluated using authoritative datasets from Johns Hopkins CSSE, WHO databases, and curated yoga pose collections. The forecasting module achieved

superior performance with ARIMA models demonstrating lowest RMSE values

compared to Prophet and LSTM approaches. The COVID checker successfully

identified five critical symptom factors (difficulty breathing, diarrhea, fever, pains,

and sore throat) through Bayesian network analysis, achieving balanced classification across demographic groups. The Yogasana recognition system reached 95% accuracy using Multi-Layer Perceptron MLP) classifiers with pose estimation via Simple HRNet and COCO pre-trained weights. The unified web application demonstrates the potential for integrated health platforms that combine predictive analytics with personalized wellness interventions, offering a scalable framework for future pandemic preparedness and holistic health management.