Abstract—In this paper we have proposed a hybrid model of Neural Networks which will predict the optimum cost for keywords in online advertising of Cost Per Click (CPC) model. Neural network is best known for finding hidden relationship, pattern or prediction of unknown variables; we have combined machine learning approach of Pattern Recognition Model and Error Correction Model in Neural Networks to calculate appropriate values for CPC. The first phase, pattern recognition model, stores keyword and its respective derived numeric value, we call it as efficiency of a keyword. The second phase, error correction model, considers all numeric features plus derived variable from first phase for prediction of CPC. In error correction model, with supervised learning weights are adjusted depending upon difference between actual value and calculated value.
Index Terms—Optimization, neural network, pay per click, prediction machine learning.
Pradeep Chaudhari is with the Tibco Software, Pune, India (e-mail: chaudharipradeep9@gmail.com). Ravindra Dingankar is with John Deere Technology Center India, Pune, India (e-mail: chaudharipradeep9@gmail.com). Roshan Chaudhari is with the Imagination Technology, Pune, India (e-mail: rgc183@gmail.com).
[PDF]
Cite:Pradeep Chaudhari, Ravindra Dingankar, and Roshan Chaudhari, "Prediction of CPC Using Neural Networks for Minimization of Cost," International Journal of Computer Theory and Engineering vol. 5, no. 4, pp. 650-652, 2013.
