使贝叶斯推理的定性数据量化

如何解决使贝叶斯推理的定性数据量化

我有一个场景，其中提出了 5 个问题。响应是二进制 1 - 是和 0 - 否。我根据它们的总和创建了一个关于最终结果的可能性的预测类。

library(dplyr)
library(ggplot2)
library(plotly)
library(purrr)
library(ggjoy)
library(forcats)

Question1 <- "Does source A state XYZ?"
Question2 <- "Does source B state XYZ?"
Question3 <- "Does source A state QRS?"
Question4 <- "Does source B state QRS?"
Question5 <- "Does source C state MNO?"
Conclusion <- "How likely is situation X to happen?"

Date <- seq.Date(from = as.Date("2020-01-01"),to = as.Date("2020-5-29"),by = "day")
Q1 <- sample(0:1,length(Date),replace = T,prob = c(0.4,0.6))
Q2 <- sample(0:1,prob = c(0.5,0.5))
Q3 <- sample(0:1,0.5))
Q4 <- sample(0:1,0.5))
Q5 <- sample(0:1,0.6))

sample_df <- data.frame(Q1,Q2,Q3,Q4,Q5)
sample_df$Sum <- rowSums(sample_df)
sample_df %>% mutate(Conclusion = case_when(Sum == 5 ~ "Almost Certain",Sum == 4 ~ "Very Likely",Sum == 3 ~ "Likely",Sum == 2 ~ "Unlikely",Sum == 1 ~ "Very Unlikely",Sum == 0 ~ "Remote")) -> sample_df

sample_df <- cbind(Date,sample_df) %>% arrange(Date)

我想使用贝叶斯推理，但不确定更新我对最终结论的预测的最佳方法。对此类数据使用 BI 有什么建议吗？目标是使用定性数据和贝叶斯来更新结论。