The production of movies requires knowledge in cinematography, in particular on techniques such as framing, camera placement and editing. For a range of 3D applications such as film prototyping and interactive narration systems, the automatization of the editing process appears a useful way of providing shot sequences that convey the story and are correct with respect to cinematographic conventions. Existing systems have used limited amounts of carefully chosen cameras. This ensures that the edit respects cinematographic conventions, but lacks variability in directorial style. In this report, we introduce a model of cinematographic editing evaluating shots and cuts in a generic fashion. This casts the problem as independent from camera generation and allows the use of all kinds of shots and cuts. Furthermore, we propose the use of maching learning methods for tuning the parameters of the model with examples.