Machine learning models are generally more accurate than statistical models. The level of interpretability differs for both the models. These arise from a number of differences, some of them being :
- Statistical modeling usually works with parametric approaches. That means we specify the number of parameters upfront whereas, in machine learning, we often use nonparametric approaches, which means that we don’t specify the structure of the model initially.
- In statistical modeling, hypothesis testing is taken into consideration along with other testing methods before implementing the model. As compared to ML models where we typically don’t make any assumptions like non-collinearity, normally distributed residuals, etc
You can check this blog out for understanding Modelling more efficiently.