Revolutionizing R&D: New Model Predicts Patent Applications with Precision
The article analyzes how patents relate to research spending using different statistical models. They looked at data from 346 companies over 10 years. They used the Poisson model and the negative binomial model to see which one fit the data better. The negative binomial model with fixed effects was found to be the best. They also used GMM to improve the accuracy of their results.