This model classifies sentiment of the scientific text based on it's context, i.e text from scientific journals to negative (n), positive (p) and neutrals (o).