Gerardo Josue Rivera Santos

Gerardo Josue Rivera's project is about the development of different models that can represent the conditions of the thermosphere. He combines the Subjects of Space Weather and Machine Learning to be able to accomplish this objective. The focus has been on being able to represent the solar effects of the different gases in the thermosphere during a full solar cycle, and this is accomplished by looking into using empirical models. The one he has been using has been the Mass Spectrometer and Incoherent Scatter (MSIS). A set of data is obtained from MSIS which is divided into different sets for training to feed into a convolutional neural network, then validate, and test the model. The model’s performance is revised by using metrics such as the Pearson Correlation Coefficient to check the spatial correlation of the data when it is reduced in its lower dimensionality form and Mean Absolute Error for reconstruction when the prediction is developed.