PolRam: A Machine Learning Based Web-Solution to Forecast and Alleviate Pollen Outbreaks

In the sphere of ecological health, the escalating prevalence of airborne allergens, particularly pollen outbreaks, poses a substantial challenge to public well-being. Traditionally, a Hirst-Trap device was used to determine the pollen count in the air, which was labor-intensive and time-consuming. To address this issue, a web application named "PolRam" is designed to forecast pollen outbreaks in the environment. The proposed application works on meteorological datasets such as minimum and maximum temperature, wind speed, humidity, precipitation, and pollen count using Machine Learning (ML) to predict pollen outbreaks in geographical locations. Using this application, users can get an early prediction of allergy forecast at a particular location. Further, an ML model, i.e., MIMO-TCN (Multiple Input Multiple Output- Temporal Convolutional Network), along with an advanced Fire-Fly algorithm, is used to predict the model performance. As a result, the proposed model attained 90 to 95% accuracy in predicting the pollen outbreak in distinct regions of India.