next up previous contents home.gif
Next: Other Algorithms Up: Modules for Special Data Previous: Data With Logical Constraints   Contents

Which modules work together

Not all the distributional models, modules for special types of data, and other features such as the dynamic library and rejection sampler can all be implemented simultaneously on the same dataset. Generally the more common modules can work with most other features, but in more specialized models for uncommon types of data, only those features one would expect to need in usual circumstances have been implemented. The chart in table 1 reflects which features can operate simultaneously.


Table 1: A chart explaining which features can work together simultaneously. Two features which meet at a darkened bullet $ \bullet $, can work in combination, those with a hollow bullet $ \circ $, do not presently work together but are planned to in future releases. An empty space means the two features can not run on the same dataset at the same time.
Module 1 2 3 4 5 6 7 8
1 Ridge Prior Model $ \bullet $ $ \bullet $ $ \circ $ $ \bullet $ $ \bullet $ $ \bullet $ $ \bullet $ $ \bullet $
2 Conditional Model $ \bullet $ $ \bullet $   $ \circ $ $ \bullet $ $ \bullet $ $ \bullet $ $ \circ $
3 T-distribtd Model $ \circ $   $ \bullet $   $ \bullet $ $ \bullet $ $ \bullet $ $ \bullet $
4 Dynamic Library $ \bullet $ $ \circ $   $ \bullet $ $ \bullet $ $ \bullet $ $ \bullet $ $ \bullet $
5 Discrete Variables $ \bullet $ $ \bullet $ $ \bullet $ $ \bullet $ $ \bullet $ $ \bullet $ $ \bullet $ $ \bullet $
6 T.S.C.S. Variables $ \bullet $ $ \bullet $ $ \bullet $ $ \bullet $ $ \bullet $ $ \bullet $   $ \circ $
7 Eff. Vote Variables $ \bullet $ $ \bullet $ $ \bullet $ $ \bullet $ $ \bullet $   $ \bullet $ $ \bullet $
8 Rejection Sampler $ \bullet $ $ \circ $ $ \bullet $ $ \bullet $ $ \bullet $ $ \circ $ $ \bullet $ $ \bullet $


As the legend explains, if two features can work together they should meet in the chart at a darkened bullet. For example, the $ t$-distributed model works with discrete variables, is planned to work with the ridge prior in future versions, and does not work with the dynamic library. Furthermore, any three or more features that can each work together pairwise can work together all at the same time. If the features needed for your dataset do not work together, first check that you have the most recent version of $ {\mathfrak{A}melia}$ and then contact one of the project members and we may implement it for you.


next up previous contents home.gif
Next: Other Algorithms Up: Modules for Special Data Previous: Data With Logical Constraints   Contents
Gary King 2003-07-25