Multilayer perceptrons neural network based web spam detection application

Web spam detection is a crucial task due to its devastationtowards Web search engines and global cost of billiondollars annually. For these reasons, a multilayeredperceptrons (MLP) neural network is presented in this paperto improve the Web spam detection accuracy. MLP neuralnetwork is used for Web...

Full description

Bibliographic Details
Main Authors: Goh, K.L., Singh, Ashutosh Kumar, Lim, King Hann
Other Authors: IEEE
Format: Conference Paper
Published: IEEE 2013
Subjects:
Online Access:http://hdl.handle.net/20.500.11937/40496
Description
Summary:Web spam detection is a crucial task due to its devastationtowards Web search engines and global cost of billiondollars annually. For these reasons, a multilayeredperceptrons (MLP) neural network is presented in this paperto improve the Web spam detection accuracy. MLP neuralnetwork is used for Web spam classification due to itsflexible structure and non-linearity transformation toaccommodate latest Web spam patterns. An intensiveinvestigation is carried out to obtain an optimal number ofhidden neurons. Both Web spam link-based and contentbasedfeatures are fed into MLP network for classification.Two benchmarking datasets – WEBSPAM-UK2006 andWEBSPAM-UK2007 are used to evaluate the performanceof the proposed classifier. The overall performance iscompared with the state of the art support vector machine(SVM) which is widely used to combat Web spam. Theexperiments have shown that MLP network outperformsSVM up to 14.02% on former dataset and up to 3.53% onlater dataset.