SGX and SG Com automatically select expressions at timings dictated by changes in the voice. The selection of expression is determined by the current behavior mode and tone: an expression will be selected randomly from the group of expressions under that behavior mode and tone. See Behavior Modes.
Although the selection is random, it is deterministic: running the analyis again on the same audio file will yield the same sequence of expressions, provided nothing has changed in the Processing Configuration.
Q: Can you weight how frequently an expression gets picked?
A: No, you cannot weight the probability of an expression. However, one way to achieve a similar result would be to duplicate the expression(s) that you want more heavily weighted so it takes up proportionally more of the mode.