Censoring approach and rounded zeros in compositional data
Project Description:
Compositional data are data which sums to a constant. It is very common in areas like geology,medical sciences, ecology, sociology, econometrics and psychology. Handling these data usually involve logratio transform. Special techniques are required to handle zeros in these data. In this project, new techniques are developed to handle the zeros more efficiently.
Project Duration:
01/2011 to 07/2018
Project Significance:
The method developed will be very useful in handling census data, voting data, food constituents data, blood compositions data, survey data etc.
Results Achieved:
A scheme had been developed and applied successfully to some datasets.
Remarks:
HPC is required for the MCMC/ HMC computation.