/semi-supervised-learning

Active learning, Passive Learning and Monte Carlo Simulations

Primary LanguageJupyter Notebook

Semi Supervised Learning using Active learning, Passive Learning and Monte Carlo Simulations

The data set can be found here

  • Objective is to implement Active Learning and Passive Learning using L1 penalized SVMs (LinearSVC)
  • Active Learning implemented by finding a sample of 10 data points closest to the margin of the SVM (As these have the most contradiction)
  • Passive Learning implemented by finding the sample randomly
  • The Monte Carlo Simulation shows how the error descends quickly for active learners as compared to passive learners