Sunspot Predictor (ML Model)

With the help of AI and Machine Learning a lot of real world problems can be solved easily. One of the Important features of ML is the machine can be retrained to do it’s job with more precision. AI and ML are generic we can apply them to any field. On surveying nearly 200 students it was evident that they were curious about ML models and how they can help in exploration and understanding of Sun.
So here is a particular application of ML to Heliophysics. Our team found a dataset from kaggle which had the data about monthly sunspots observed during the period 1800 – 2019. We trained a LSTM based Neural Network on the data and tried to forecast the results in the test dataset the model’s performance was astonishing. This could give the people a lot of information on how data from solar expeditions could look like and what can we do with those data. From this model we can understand about the solar cycle.
If future we aim to Predict and forecast the geomagnetic disturbances caused due to solar wind which would help us to understand a lot of natural phenomena like solar flares, Auroras and most importantly in terraforming Mars and etc…