misvm.md
In mildsvm: Multiple-Instance Learning with Support Vector Machines

`misvm()` value returns make sense

Code
  models <- list(`xy-heur` = run_misvm(method = "heuristic"), `xy-mip` = run_misvm(
    method = "mip"), `xy-qp` = run_misvm(method = "qp-heuristic"), formula = misvm(
    mi(bag_label, bag_name) ~ X1_mean + X2_mean, method = "heuristic", data = df1),
  mi_df = misvm(as_mi_df(df1, instance_label = NULL)), mildata = misvm(mil_data),
  `no-scale-heur` = run_misvm(method = "heuristic", control = list(scale = FALSE)),
  `no-scale-mip` = run_misvm(method = "mip", control = list(scale = FALSE)),
  `no-scale-qp` = run_misvm(method = "qp-heuristic", control = list(scale = FALSE)),
  kfm_fit = misvm(mi(bag_label, bag_name) ~ X1_mean + X2_mean, data = df1,
  method = "mip", control = list(kernel = "radial")), `no-weights-heur` = run_misvm(
    method = "heuristic", weights = FALSE), `no-weights-mildata` = misvm(mil_data)) %>%
    suppressWarnings() %>% suppressMessages()
  print(lapply(models, names))
Output
  $`xy-heur`
   [1] "svm_fit"   "call_type" "x"         "features"  "levels"    "cost"     
   [7] "weights"   "kernel"    "repr_inst" "n_step"    "x_scale"

  $`xy-mip`
  [1] "gurobi_fit" "call_type"  "features"   "levels"     "cost"      
  [6] "weights"    "kernel"     "x_scale"

  $`xy-qp`
   [1] "gurobi_fit" "call_type"  "features"   "levels"     "cost"      
   [6] "weights"    "kernel"     "repr_inst"  "n_step"     "x_scale"

  $formula
   [1] "svm_fit"   "call_type" "x"         "features"  "levels"    "cost"     
   [7] "weights"   "kernel"    "repr_inst" "n_step"    "x_scale"   "formula"  
  [13] "bag_name"

  $mi_df
   [1] "svm_fit"   "call_type" "x"         "features"  "levels"    "cost"     
   [7] "weights"   "kernel"    "repr_inst" "n_step"    "x_scale"   "bag_name"

  $mildata
   [1] "svm_fit"       "call_type"     "x"             "features"     
   [5] "levels"        "cost"          "weights"       "kernel"       
   [9] "repr_inst"     "n_step"        "x_scale"       "bag_name"     
  [13] "instance_name" "summary_fns"   "summary_cor"

  $`no-scale-heur`
  [1] "svm_fit"   "call_type" "features"  "levels"    "cost"      "weights"  
  [7] "kernel"    "repr_inst" "n_step"

  $`no-scale-mip`
  [1] "gurobi_fit" "call_type"  "features"   "levels"     "cost"      
  [6] "weights"    "kernel"

  $`no-scale-qp`
  [1] "gurobi_fit" "call_type"  "features"   "levels"     "cost"      
  [6] "weights"    "kernel"     "repr_inst"  "n_step"

  $kfm_fit
   [1] "gurobi_fit"   "kfm_fit"      "call_type"    "features"     "levels"      
   [6] "cost"         "weights"      "kernel"       "kernel_param" "x_scale"     
  [11] "formula"      "bag_name"

  $`no-weights-heur`
   [1] "svm_fit"   "call_type" "x"         "features"  "levels"    "cost"     
   [7] "kernel"    "repr_inst" "n_step"    "x_scale"

  $`no-weights-mildata`
   [1] "svm_fit"       "call_type"     "x"             "features"     
   [5] "levels"        "cost"          "weights"       "kernel"       
   [9] "repr_inst"     "n_step"        "x_scale"       "bag_name"     
  [13] "instance_name" "summary_fns"   "summary_cor"

Code
  print(models)
Output
  $`xy-heur`
  An misvm object called with misvm.default

  Parameters: 
    method: heuristic 
    kernel: linear  
    cost: 1 
    scale: TRUE 
    weights: ('0' = 0.375, '1' = 1)

  Model info: 
    Features: chr [1:120] "X1_0.05" "X1_0.15" "X1_0.25" "X1_0.35" ...
    Number of iterations: 2


  $`xy-mip`
  An misvm object called with misvm.default

  Parameters: 
    method: mip 
    kernel: linear  
    cost: 1 
    scale: TRUE 
    weights: ('0' = 0.375, '1' = 1)

  Model info: 
    Features: chr [1:120] "X1_0.05" "X1_0.15" "X1_0.25" "X1_0.35" ...
    Gap to optimality: 0


  $`xy-qp`
  An misvm object called with misvm.default

  Parameters: 
    method: qp-heuristic 
    kernel: linear  
    cost: 1 
    scale: TRUE 
    weights: ('0' = 0.375, '1' = 1)

  Model info: 
    Features: chr [1:120] "X1_0.05" "X1_0.15" "X1_0.25" "X1_0.35" ...
    Number of iterations: 2


  $formula
  An misvm object called with misvm.formula

  Parameters: 
    method: heuristic 
    kernel: linear  
    cost: 1 
    scale: TRUE 
    weights: ('0' = 0.375, '1' = 1)

  Model info: 
    Features: chr [1:2] "X1_mean" "X2_mean"
    Number of iterations: 2


  $mi_df
  An misvm object called with misvm.mi_df

  Parameters: 
    method: heuristic 
    kernel: linear  
    cost: 1 
    scale: TRUE 
    weights: ('0' = 0.375, '1' = 1)

  Model info: 
    Features: chr [1:120] "X1_0.05" "X1_0.15" "X1_0.25" "X1_0.35" ...
    Number of iterations: 2


  $mildata
  An misvm object called with misvm.mild_df

  Parameters: 
    method: heuristic 
    kernel: linear  
    cost: 1 
    scale: TRUE 
    weights: ('0' = 0.375, '1' = 1)

  Model info: 
    Features: chr [1:20] "X1_mean" "X2_mean" "X3_mean" "X4_mean" "X5_mean" ...
    Number of iterations: 2


  $`no-scale-heur`
  An misvm object called with misvm.default

  Parameters: 
    method: heuristic 
    kernel: linear  
    cost: 1 
    scale: FALSE 
    weights: ('0' = 0.375, '1' = 1)

  Model info: 
    Features: chr [1:120] "X1_0.05" "X1_0.15" "X1_0.25" "X1_0.35" ...
    Number of iterations: 3


  $`no-scale-mip`
  An misvm object called with misvm.default

  Parameters: 
    method: mip 
    kernel: linear  
    cost: 1 
    scale: FALSE 
    weights: ('0' = 0.375, '1' = 1)

  Model info: 
    Features: chr [1:120] "X1_0.05" "X1_0.15" "X1_0.25" "X1_0.35" ...
    Gap to optimality: 0


  $`no-scale-qp`
  An misvm object called with misvm.default

  Parameters: 
    method: qp-heuristic 
    kernel: linear  
    cost: 1 
    scale: FALSE 
    weights: ('0' = 0.375, '1' = 1)

  Model info: 
    Features: chr [1:120] "X1_0.05" "X1_0.15" "X1_0.25" "X1_0.35" ...
    Number of iterations: 3


  $kfm_fit
  An misvm object called with misvm.formula

  Parameters: 
    method: mip 
    kernel: radial  (sigma = 0.5) 
    cost: 1 
    scale: TRUE 
    weights: ('0' = 0.375, '1' = 1)

  Model info: 
    Features: chr [1:2] "X1_mean" "X2_mean"
    Gap to optimality: 0


  $`no-weights-heur`
  An misvm object called with misvm.default

  Parameters: 
    method: heuristic 
    kernel: linear  
    cost: 1 
    scale: TRUE 
    weights: FALSE

  Model info: 
    Features: chr [1:120] "X1_0.05" "X1_0.15" "X1_0.25" "X1_0.35" ...
    Number of iterations: 2


  $`no-weights-mildata`
  An misvm object called with misvm.mild_df

  Parameters: 
    method: heuristic 
    kernel: linear  
    cost: 1 
    scale: TRUE 
    weights: ('0' = 0.375, '1' = 1)

  Model info: 
    Features: chr [1:20] "X1_mean" "X2_mean" "X3_mean" "X4_mean" "X5_mean" ...
    Number of iterations: 2

Ordering of data doesn't change `misvm()` results

Code
  with(df1_test, {
    pred <- predict(mdl2, df1_test, type = "raw")$.pred
    pROC::auc(classify_bags(bag_label, bag_name), classify_bags(pred, bag_name))
  })
Message <simpleMessage>
  Setting levels: control = 0, case = 1
  Setting direction: controls < cases
Output
  Area under the curve: 1

Code
  with(df1_test, {
    pred <- predict(mdl2, df1_test, type = "raw")$.pred
    pROC::auc(classify_bags(bag_label, bag_name), classify_bags(pred, bag_name))
  })
Message <simpleMessage>
  Setting levels: control = 0, case = 1
  Setting direction: controls < cases
Output
  Area under the curve: 1

Code
  with(mil_data_test, {
    pred <- predict(mdl2, mil_data_test, type = "raw")$.pred
    pROC::auc(classify_bags(bag_label, bag_name), classify_bags(pred, bag_name))
  })
Message <simpleMessage>
  Setting levels: control = 0, case = 1
  Setting direction: controls < cases
Output
  Area under the curve: 1