I have a problem where I have 5 examples of truncated bimodal distributions between different ranges of (-1.0, +1.0), (-0.18, +0.18), (-0.12, +0.12) and (-0.15, +0.15). The example distributions can be viewed here.
Please note that for all of the given example bimodal distributions, there is no zero values within the given distribution! The data distributions have values ranging between the given ranges EXCEPT of ZERO!
I am trying to create a sampling technique which can create random samples by obeying the given bimodal samples in the examples. The samples to be drawn can be of any size/shape/dimension such as 784, 5000, etc.
Any ideas how I can create such truncated bimodal data distributions between intervals 'a' and 'b' or [a, b] (both points are included)!
Thanks!