Discovery of Long Tail Keywords in Paid Search

Journal of Applied & Computational Mathematics

article

Tesiero J¹ 

[1] Principal Data Scientist Consultant, University of Maine

关键词: Clustering method;    Marin data;    Matrix yield;    Nonlinear;   

DOI  :  10.4172/2168-9679.1000315

来源: Hilaris Publisher

【 摘 要 】

The following work describes an elegant, efficient keyword clustering method to discover long tail keywords in paid search data. In keyword auctions, such words often go undiscovered as their cost in being bid to higher ranking positions is deemed too high to justify the potential of significantly added conversion revenue. By discovering clusters with low volume keywords and established, high-performing and high volume keywords, the quality of the low volume (long tail) keywords is inferred by association.After a brief introduction, the data used to train the clustering algorithm is described. Then, the data reduction process (the discovery of the most predictive features) is described. We then describe the method, followed by the results and interpretation.

【 授权许可】

Unknown