Learning Functions Using Randomized Genetic Code-Like Transformations: Probabilistic Properties and Experimentations
Issue No.08 - August (2004 vol.16)
Hillol Kargupta , IEEE
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/TKDE.2004.27
<p><b>Abstract</b>—Inductive learning of nonlinear functions plays an important role in constructing predictive models and classifiers from data. This paper explores a novel randomized approach to construct linear representations of nonlinear functions proposed elsewhere [CHECK END OF SENTENCE], [CHECK END OF SENTENCE]. This approach makes use of randomized codebooks, called the Genetic Code-Like Transformations (GCTs) for constructing an approximately linear representation of a nonlinear target function. This paper first derives some of the results presented elsewhere [CHECK END OF SENTENCE] in a more general context. Next, it investigates different probabilistic and limit properties of GCTs. It also presents several experimental results to demonstrate the potential of this approach.</p>
Inductive function learning, genetic code-like transformations, representation construction, randomized transformations.
Hillol Kargupta, Rajeev Ayyagari, Samiran Ghosh, "Learning Functions Using Randomized Genetic Code-Like Transformations: Probabilistic Properties and Experimentations", IEEE Transactions on Knowledge & Data Engineering, vol.16, no. 8, pp. 894-908, August 2004, doi:10.1109/TKDE.2004.27