|The COSNet R package
Next : Software and documentation download
Overview of the COSNet R Package
- COSNet (COst Sensitive neural Network) [1, 2]
is a novel method to learn node labels biological networks with a high
prevalence of negative instances. Examples of this context are the
automated prediction of protein functions, the gene disease
prioritization and the drug reposition problem.
- COSNet is based on a
cost-sensitive family of parametrized Hopfield networks, whose
characteristics can be summarized as follows:
- Class labels and neuron states are conceptually separated. In this way a
class of Hopfield networks is introduced, having as parameters the
values of neuron states and the neuron thresholds.
- The parameters
of the network are learned from the data through an efficient supervised
algorithm, in order to take into account the unbalance between positive
and negative node labels.
- The dynamics of the network is
restricted to its unlabeled part, preserving the minimization of the
overall objective function and the a priori information and
significantly reducing the time complexity of the learning algorithm.
COSNet package provides methods to support all the abovementioned
steps, and methods for cross validating the algorithm and for
stratified and not stratified partition of the input data. See section Main functionalities of COSNet.
The instruction for downloading the R software and the documentation
(the reference manual and the vignette in pdf format) are available in
the section Software and documentation download. To install the software go to section Software installation, whereas for an example of COSNet library usage go to section A step-by-step application of COSNet to protein function prediction.
Frasca, M., Bertoni, A., Re, M. and Valentini, G.
A neural network algorithm for semi-supervised node label learning
from unbalanced data. Neural Networks, 43, 84-98, 2013.
A. Bertoni, M. Frasca, G.
Valentini COSNet: a Cost Sensitive Neural Network for Semi-supervised Learning in Graphs. In:"Machine Learning and Knowledge Discovery in
Databases". European Conference, ECML PKDD 2011, Athens,
Greece, Proceedings, Part I, Lecture Notes in Artificial Intelligence, vol.
6911, pp.219-234, Springer, 2011.