Confidence regions in ternary diagrams based on the power-divergence statistics
Watson and Nguyen (1985) and Watson (1987) consider the problem of plotting confidence regions in a ternary diagram for a trinomial probability vector II based on Pearson. Their results are extended to the power-divergence family of statistics resulting in confidence regions of diverse shapes and sizes. The members of the family with the most accurate coverage probabilities are A=2/3 and A=1/2
Please refer to publisher version or contact your library.