This merge request contains major changes. First, the new functions of the two stage classifer are added to the POSEIDON library. Second, major changes were made in the prototype.R file. Third, using the code from prototype.R as well as new code I created a detailled vignette to describe how to fit models, make inference and perform classification based on non-sequential and sequential (two-stage and greedy/exhaustive). Importantly, the function for the two stage classifier was adapted in a way that allows for flexibel classification. I even could add another examples covering measusurements (at baseline) of MRI and MMSE in the first stage and optional follow-up MRI and MMSE measures in the second stage. It is only done for one subject and not over the whole sample, nevertheless we can still mention it shortly in the paper when we also make a point about the impact because we created an own library.