Page History

Configuration

• Parameter
  summary Model to use for classification, by default it always uses the latest, but can be configure to use one in specific default --LATEST-- name Model required true

• Parameter
  summary Probability threshold. Will only tag sentences that match better or equal to the value specified. default .95 name Minimum Probability type double required true

• Parameter
  summary Indicates which tags must be first recognized in order to do the classification and training name Normalize Tags type string array

  
  
  • This helps to reduce the noise in the text, for example reducing every set of numbers to {number}

• Parameter
  summary The amount of training data which represents the positive samples default 0.5 name Positive Sample Ration type double

  
  
  • If there are 5000 positive samples, that 5000 represents 60% (with a PSR of 0.6) of the total training data, where the other 40% will be negative samples, around 3333 negative samples

Training a Model

Click on the  button which will popup the "Start Training Run" dialog

• Parameter
  summary Select Selects the Dataset to use as training data name Datasets type boolean required true

• Parameter
  default 200 name Iterations type integer required true

• Parameter
  default 5 name Cut Off type integer required true

• Parameter
  default 2 name Threads type integer required true

• Parameter
  default BoW name Feature Selection required true

  • BoW (Bag of Words)
  • N-Gram

• Parameter
  default MAXENT_QN name Algorithm required true

  • Available algorithms

    • MAXENT_QN

    • MAXENT

    • NAIVEBAYES

    • PERCEPTRON

• Parameter
  default 0.1 name L1Cost type double required true

• Parameter
  default 0.1 name L2Cost type double required true

• Parameter
  default 15 name Number of Updates type integer required true

• Parameter
  default 30000 name Max FctEval type integer required true

General Settings