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

Re-ordering data doesn't reduce performance

Code
  auc1 <- with(mil_df_test, pROC::auc(response = classify_bags(bag_label,
    bag_name), predictor = classify_bags(pred1$.pred, bag_name)))
Message <simpleMessage>
  Setting levels: control = 0, case = 1
  Setting direction: controls < cases
Code
  auc2 <- with(mil_df_test, pROC::auc(response = classify_bags(bag_label,
    bag_name), predictor = classify_bags(pred2$.pred, bag_name)))
Message <simpleMessage>
  Setting levels: control = 0, case = 1
  Setting direction: controls < cases
Code
  auc1
Output
  Area under the curve: 1
Code
  auc2
Output
  Area under the curve: 1
Code
  eps <- 0.05

`smm()` value returns make sense

Code
  models <- list(xy = smm(x, y, instances), formula = smm(y ~ x1 + x2 + x3, data = df),
  mildata = smm(mil_df), `no-scale-xy` = smm(x, y, instances, control = list(
    scale = FALSE)), `no-scale-mildata` = smm(mil_df, control = list(scale = FALSE)),
  `no-weights-xy` = smm(x, y, instances, weights = FALSE), `no-weights-mildata` = smm(
    mil_df, weights = FALSE)) %>% suppressWarnings() %>% suppressMessages()
  print(lapply(models, names))
Output
  $xy
   [1] "ksvm_fit"     "call_type"    "x"            "features"     "levels"      
   [6] "cost"         "sigma"        "weights"      "kernel"       "kernel_param"
  [11] "x_scale"

  $formula
   [1] "ksvm_fit"      "call_type"     "x"             "features"     
   [5] "levels"        "cost"          "sigma"         "weights"      
   [9] "kernel"        "kernel_param"  "x_scale"       "formula"      
  [13] "instance_name"

  $mildata
   [1] "ksvm_fit"      "call_type"     "x"             "features"     
   [5] "levels"        "cost"          "sigma"         "weights"      
   [9] "kernel"        "kernel_param"  "x_scale"       "bag_name"     
  [13] "instance_name"

  $`no-scale-xy`
   [1] "ksvm_fit"     "call_type"    "x"            "features"     "levels"      
   [6] "cost"         "sigma"        "weights"      "kernel"       "kernel_param"

  $`no-scale-mildata`
   [1] "ksvm_fit"      "call_type"     "x"             "features"     
   [5] "levels"        "cost"          "sigma"         "weights"      
   [9] "kernel"        "kernel_param"  "bag_name"      "instance_name"

  $`no-weights-xy`
   [1] "ksvm_fit"     "call_type"    "x"            "features"     "levels"      
   [6] "cost"         "sigma"        "kernel"       "kernel_param" "x_scale"

  $`no-weights-mildata`
   [1] "ksvm_fit"      "call_type"     "x"             "features"     
   [5] "levels"        "cost"          "sigma"         "kernel"       
   [9] "kernel_param"  "x_scale"       "bag_name"      "instance_name"

Code
  print(models)
Output
  $xy
  A smm object called with smm.default

  Parameters: 
    kernel: kme w/ radial  (sigma = 0.3333333) 
    cost: 1 
    scale: TRUE 
    weights: ('-1' = 1, '1' = 1)

  Model info: 
    Features: chr [1:3] "x1" "x2" "x3"


  $formula
  A smm object called with smm.formula

  Parameters: 
    kernel: kme w/ radial  (sigma = 0.3333333) 
    cost: 1 
    scale: TRUE 
    weights: ('-1' = 1, '1' = 1)

  Model info: 
    Features: chr [1:3] "x1" "x2" "x3"


  $mildata
  A smm object called with smm.mild_df

  Parameters: 
    kernel: kme w/ radial  (sigma = 0.1) 
    cost: 1 
    scale: TRUE 
    weights: ('0' = 1.5, '1' = 1)

  Model info: 
    Features: chr [1:10] "X1" "X2" "X3" "X4" "X5" "X6" "X7" "X8" "X9" ...


  $`no-scale-xy`
  A smm object called with smm.default

  Parameters: 
    kernel: kme w/ radial  (sigma = 0.3333333) 
    cost: 1 
    scale: FALSE 
    weights: ('-1' = 1, '1' = 1)

  Model info: 
    Features: chr [1:3] "x1" "x2" "x3"


  $`no-scale-mildata`
  A smm object called with smm.mild_df

  Parameters: 
    kernel: kme w/ radial  (sigma = 0.1) 
    cost: 1 
    scale: FALSE 
    weights: ('0' = 1.5, '1' = 1)

  Model info: 
    Features: chr [1:10] "X1" "X2" "X3" "X4" "X5" "X6" "X7" "X8" "X9" ...


  $`no-weights-xy`
  A smm object called with smm.default

  Parameters: 
    kernel: kme w/ radial  (sigma = 0.3333333) 
    cost: 1 
    scale: TRUE 
    weights: FALSE

  Model info: 
    Features: chr [1:3] "x1" "x2" "x3"


  $`no-weights-mildata`
  A smm object called with smm.mild_df

  Parameters: 
    kernel: kme w/ radial  (sigma = 0.1) 
    cost: 1 
    scale: TRUE 
    weights: FALSE

  Model info: 
    Features: chr [1:10] "X1" "X2" "X3" "X4" "X5" "X6" "X7" "X8" "X9" ...