Example Pop vs Soda vs Coke
Define the prior probability model.
prior <- c("midwest" = 0.21, "northeast" = 0.17, "south" = 0.38, "west" = 0.24)
Prior probability model |
region |
midwest |
northeast |
south |
west |
probability |
21.0% |
17.0% |
38.0% |
24.0% |
likelihood <- pop_vs_soda |>
tabyl(pop, region) |>
adorn_percentages("col") |>
filter(pop) |>
select(-pop)
Data model (likelihoods) |
region |
midwest |
northeast |
south |
west |
likelihood |
64.5% |
27.3% |
7.9% |
29.4% |
posterior <- prior * likelihood / sum(prior * likelihood)
Posterior porbability model |
|
midwest |
northeast |
south |
west |
prior |
21.0% |
17.0% |
38.0% |
24.0% |
likelihood |
64.5% |
27.3% |
7.9% |
29.4% |
posterior |
47.9% |
16.4% |
10.7% |
25.0% |
We can summarize the key observations from this analysis is the form of the below graph.