test.Rmd

---
title: "R Notebook"
output: none
---

```{r}
library("ai")

  submat <- c('BLOSUM45', 'BLOSUM50', 'BLOSUM62', 'BLOSUM80', 'BLOSUM100', 'PAM30', 'PAM40', 'PAM70', 'PAM120', 'PAM250')
  sheet_sup4 <- c("Van Allen et al.", "Snyder et al.", "Rizvi et al.")
  a <- 1

  # Load data
    input_data <- read.csv2(paste0("/Users/urniaz/Desktop/kmeRs/Paper/data_new/", sheet_sup4[a],'csv'))
    input_data <- input_data[ ,c(-1, -2, -3, -5, -6)]
    input_data <- data.frame(apply(input_data[, ], 2, function(x) as.numeric(as.character(x))))

  # Collect stats
  res_frame <- data.frame(matrix(ncol = 5, nrow = 0))
  
for (i in 1:1) { # How much tests run?
  # Data
    data <- ai::prodata(input_data, "Status") # <- n randomizacji tutaj i na tcyh samych zbirach danych puscici wszytskie modele i kombinacje z PAM i BLOSUM 
    
  # Model A <- tylko raz
    config <- ai::config(formula = "Status ~ A + R", k = 3)
    A <- ai::model(data = data, type = c("lm", "rlm", "knn", "knn1", "ctree"), config = config)
    
# For (e in 1:length (submat)) !!!!
    
   # Model B <- wiele razy bo kombinacje pam blisum 
    config <- ai::config(formula = "Status ~ A + R", k = 3)
    B <- ai::model(data = data, type = c("lm", "rlm", "knn", "knn1", "ctree"), config = config)
    
  # Stats
    A <- ai::stats_model(A)
    B <- ai::stats_model(B)
    
  # Compare
  st <- ai::stats_compare_models(A, B)
  
  # Collect results
  res_frame <- rbind(res_frame, data.frame(loop = i, st))
   
}
```