Available algorithms

A list of implemented algorithms (as web services) provided by opentox.ntua.gr is available at http://opentox.ntua.gr:8080/algorithm.

The implemented algorithms so far include:

  1. Fast RBF NN: Fast-RBF-NN is a training algorithm for Radial Basis Function Neural Networks. The algorithm is based on the subtractive clustering technique and has a number of advantages compared to the traditional learning algorithms including faster training times and more accurate predictions
  2. Scaling: All features are scaled within a given interval
  3. Support Vector Machines: Algorithm for training regression models using the Support Vector Machine Learning Algorithm. The training is based on the Weka implementation of SVM and specifically the class weka.classifiers.functions.SVMreg.
  4. Leverage DoA: The well known leverages algorithm for the estimation of a model's applicability domain
  5. Multiple Linear Regression: Training algorithm for multiple linear regression models. Applies on datasets which contain exclusively numeric data entries. The algorithm is an implementation of LinearRegression of Weka.
  6. Cleanup: Removes from the dataset all attributes of certain types. This is useful in many cases, when for example one needs to train a model but have a clean dataset that doesn't include string values