5_ede_classification_y2.yaml

Connector:
#  PREndpoint: 194.102.62.155 #hal720m.sage.ieat.ro
  Dask:
    SchedulerEndpoint: local # if not local add DASK schedueler endpoint
    Scale: 3 # Number of workers if local otherwise ignored
    SchedulerPort: 8787 # This is the default point
    EnforceCheck: False # Irrelevant for local
  MPort: 9200 # Moitoring port
  KafkaEndpoint: 10.9.8.136
  KafkaPort: 9092
  KafkaTopic: edetopic
#  Query: { "query": 'node_disk_written_bytes_total[5m]'} # Query for specific metrics
  Query: {"query": '{__name__=~"node.+"}[1m]'}
  MetricsInterval: "1m" # Metrics datapoint interval definition
  QSize: 0
  Index: time
  QDelay: "10s" # Polling period for metrics fetching
  Local: /Users/Gabriel/Dropbox/Research/ASPIDE/Datasets/ECI Chaos/Distributed Phase 1/finalized/single_node/training/df_anomaly.csv # Define the path to the local file for training

Mode:
  Training: True
  Validate: False
  Detect: False


Filter:
  Fillna: True # fill none values with 0
  Dropna: True # delete columns woth none values

Augmentation:
  Scaler: # if not used set to false
    StandardScaler:   # All scalers from scikitlearn
      copy: True

#  Classification example
#Training:
#  Type: classification
#  Method: !!python/object/apply:sklearn.ensemble.AdaBoostClassifier  # DONT forger ../apply
#    _sklearn_version: '0.24.2'
#    n_estimators: 100
#    learning_rate: 1
#    algorithm: SAMME.R
#  Target: target
#  Export: classification_2
#  ValidRatio: 0.2
#  TrainScore: True # expensive if set to false only test scores are computed
#  ReturnEstimators: True
#  CV:
#    Type: StratifiedKFold  # user defined all from sklearn, if int than used standard
#    Params:
#      n_splits: 5
#      shuffle: True
#      random_state: 5
#  Scorers:
#    Scorer_list:
#      - Scorer:
#          Scorer_name: F1_weighted
#          skScorer: f1_weighted
#      - Scorer:
#          Scorer_name: Jaccard_Index
#          skScorer: jaccard_weighted # changes in scoring sklearn, for multiclass add suffix micro, weighted or sample
#      - Scorer:
#          Scorer_name: AUC
#          skScorer: roc_auc_ovr_weighted
#    User_scorer1: balanced_accuracy_score # key is user defined, can be changed same as Scorer_name
#    Verbose: 0


#Training:
#  Type: classification
#  Method: dnn
#  MethodSettings:
#    optimizer: 'adam'
#    learning_r: 0.01
#    kernel_init: 'he_normal'
#    layer_1: 0
#    layer_2: 50
#    layer_3: 100
#    layer_0: 50
#    drop: 0.3
#    loss: 'categorical_crossentropy'
#    activation_1: 'relu'
#    out_activation: 'sigmoid'
#  Export: dnn_y2


Training:
  Type: classification
  Method: !!python/object/apply:sklearn.ensemble.RandomForestClassifier  # DONT forger ../apply
    _sklearn_version: '0.24.2'
    n_estimators: 100
    criterion: "gini"
    min_sample_split: 2
    min_sample_leaf: 1
    max_features: "log2"
    n_jobs: -1
    random_state: 42
    verbose: 1
  Target: target
  Export: classification_y2
  ValidRatio: 0.2
  TrainScore: True # expensive if set to false only test scores are computed
  ReturnEstimators: True
  CV:
    Type: StratifiedKFold  # user defined all from sklearn, if int than used standard
    Params:
      n_splits: 5
      shuffle: True
      random_state: 5
  Scorers:
    Scorer_list:
      - Scorer:
          Scorer_name: F1_weighted
          skScorer: f1_weighted
      - Scorer:
          Scorer_name: Jaccard_Index
          skScorer: jaccard_weighted # changes in scoring sklearn, for multiclass add suffix micro, weighted or sample
      - Scorer:
          Scorer_name: AUC
          skScorer: roc_auc_ovr_weighted
    User_scorer1: balanced_accuracy_score # key is user defined, can be changed same as Scorer_name

  LearningCurve:
    sizes: !!python/object/apply:numpy.core.function_base.linspace
      kwds:
        start: 0.3
        stop: 1.0
        num: 10
    scorer: f1_weighted
    n_jobs: 5

  ValidationCurve:
    param_name: n_estimators
    param_range:
      - 10
      - 20
      - 60
      - 100
      - 200
      - 600
    scoring: f1_weighted
    n_jobs: 8
  PrecisionRecallCurve: 1
  ROCAUC: 1
  RFE:
    scorer: f1_weighted
    step: 10
  DecisionBoundary: 1
  Verbose: 1

#Training:
#  Type: classification
#  Method: !!python/object/apply:xgboost.XGBClassifier  # DONT forger ../apply
#    kwds:
#      n_estimators: 100
#      max_depth: 4
#      learning_rate: 0.01
#      subsample: 0.2
#      min_child_wight: 6
#      objective: multi:softmax
#      n_jobs: -1
#      random_state: 42
#      verbosity: 1
#  Target: target
#  Export: classification_xgb_y2
#  ValidRatio: 0.2
#  TrainScore: True # expensive if set to false only test scores are computed
#  ReturnEstimators: True
#  CV:
#    Type: StratifiedKFold  # user defined all from sklearn, if int than used standard
#    Params:
#      n_splits: 5
#      shuffle: True
#      random_state: 5
#  Scorers:
#    Scorer_list:
#      - Scorer:
#          Scorer_name: F1_weighted
#          skScorer: f1_weighted
#      - Scorer:
#          Scorer_name: Jaccard_Index
#          skScorer: jaccard_weighted # changes in scoring sklearn, for multiclass add suffix micro, weighted or sample
#      - Scorer:
#          Scorer_name: AUC
#          skScorer: roc_auc_ovr_weighted
#    User_scorer1: balanced_accuracy_score # key is user defined, can be changed same as Scorer_name
#  Verbose: 1


Detect:
  Method: AdaBoostClassifier
  Type: classification
  Load: classification_1_0
  Scaler: StandardScaler # Same as for training



# Not yet Implemented
#Validation:
#  DataSource: /path/to/data # if datasource is not defined use default from data connector, last column is ground truth named "Target"
#  Treashold: 0.2 #  improvement percent
#  Models:
#    - m1
#    - m2

Misc:
  heap: 512m
  checkpoint: True
  delay: 15s
  interval: 30m
  resetindex: False
  point: False