In response to the rapid global honeybee decline, a novel lactic acid bacteria probiotic mixture was developed five years ago. Although an existing multiple linear regression (MLR) model was used to determine this mixture’s optimal dosage last year, a major problem was that the model would not minimize the sum of the residuals unless explicit training commands were given, which is why a novel MLR model was developed with continual-learning algorithms. After training, there was an insignificant difference between the novel MLR model’s predictions and experimental values. Additionally, the MLR was significantly more accurate (peak of 92.8%) than existing ML alternatives. The combination of this novel MLR and probiotic mixture can potentially reverse this alarming decline of our pollinators.
Varun Raj Madan