P value校正思想與實現

1. 爲什麼要對P-value進行校正

兩兩對比做多了，不做校正的話，會獲得很多假陽性結論。具體點這裏

2. 有哪些校正方法

常用的P-value校正方法包括：Bonferroni correction, Benjamini & Hochberg (FDR), Hommel, Benjamini & Yekutieli等。

3. P-value 校正的實現

這裏介紹得很好

Input = ("
Food               Raw.p
Blue_fish         .34
Bread             .594
Butter            .212
Carbohydrates     .384
Cereals_and_pasta .074
Dairy_products    .94
Eggs              .275
Fats              .696
Fruit             .269
Legumes           .341
Nuts              .06
Olive_oil         .008
Potatoes          .569
Processed_meat    .986
Proteins           .042
Red_meat           .251
Semi-skimmed_milk  .942
Skimmed_milk       .222
Sweets             .762
Total_calories     .001
Total_meat         .975
Vegetables         .216
White_fish         .205
White_meat         .041
Whole_milk         .039
")

Data = read.table(textConnection(Input),header=TRUE)
Data = Data[order(Data$Raw.p),]
library(FSA)
Data$Bonferroni = p.adjust(Data$Raw.p, method = "bonferroni")
Data$BH = p.adjust(Data$Raw.p, method = "BH")

有時間再詳細寫寫吧，現在僅做個記錄