Update general_evaluation.R

src/R/functions/general_evaluation.R

@@ -278,7 +278,7 @@ accuracy_and_uncertainty = function(
  sp_co = sp_co %>% filter(Freq > 0)
  #write_csv(sp_co, "./outdir/species_confusion_june23.csv")
 
- figure_1(pairs, vst)
+ figure_1(pairs, data)
 
  #plot distribution of confusion by health class
  figure_2(pairs, vst)
@@ -296,29 +296,37 @@ accuracy_and_uncertainty = function(
  pairs = readr::read_csv(pairs_fam_pt)
  pairs = cbind.data.frame(pairs, pp)
  colnames(pairs)=c("id", "obs", "pred", "i", "individualID", "obs_sp", "pred_sp")
-
+
+ ee = vst %>% select(taxonID, scientificName) %>% unique
+ ee$genus = do.call(rbind.data.frame, strsplit(ee$scientificName, split = " "))[,1]
+ ee = unique(ee)
+ ee = ee %>% select(taxonID, genus)
+ colnames(ee) = c("obs_sp", "obs_scientific")
+ pairs = left_join(pairs, ee)
+ colnames(ee) = c("pred_sp", "pred_scientific")
+ pairs = left_join(pairs, ee)
+
  pairs = pairs %>% filter(individualID %in% ids_and_genus$individualID)
- ids_and_genus$obs_scientific = do.call(rbind.data.frame, strsplit(ids_and_genus$obs_scientific, split = " "))[,1]
- ids_and_genus$pred_scientific = do.call(rbind.data.frame, strsplit(ids_and_genus$pred_scientific, split = " "))[,1]
- dm_dt = ids_and_genus
+ #ids_and_genus$obs_scientific = do.call(rbind.data.frame, strsplit(ids_and_genus$obs_scientific, split = " "))[,1]
+ #ids_and_genus$pred_scientific = do.call(rbind.data.frame, strsplit(ids_and_genus$pred_scientific, split = " "))[,1]
+ dm_dt = pairs
  genuses = c(dm_dt$obs_scientific, dm_dt$pred_scientific) %>% unique
  #lvls = unique(c(unique(genuses$obs), unique(genuses$pred)))
  dm_dt$obs = factor(dm_dt$obs_scientific, levels = unique(genuses))
  dm_dt$pred = factor(dm_dt$pred_scientific, levels = unique(genuses))
  cmdm = confusionMatrix(dm_dt$obs, dm_dt$pred)
  print(cmdm$overall)
-
+ 
  conf = as.matrix(cmdm)
  conf = conf %>% t
  conf = cbind.data.frame(rownames(conf), conf)
  write_csv(data.frame(conf), paste("./outdir/cm/Family_level_confusion_matrix.csv", sep=""))
-
-
+ 
+ 
  sp_co = cmdm$table
  sp_co = cmdm$table %>% data.frame
  sp_co = sp_co %>% filter(Freq > 0)
  write_csv(sp_co, "./outdir/genuses_confusion.csv")
-
-
+
 
 }