Statistical modeling is about applying statistics on data to find the underlying relationships by analyzing the significance of the variables. It is the formalization of relationships between the variables in the form of mathematical equations.
As it is only a model it predicts the output with accuracy of 85 percent and having 90 percent confidence about it. For this purpose, various diagnostics of parameters can be performed, like p-value for example.
The model will need a 70 percent - 30 percent split of data in order to create training and testing data so that it will be developed on training data and tested on testing data.