‘Learning to Rank’ Text Search Engine Platform for Internal Wikis
A large number of companies maintain an internal workplace wiki to document specific goals or processes pertaining to different projects. These wikis can grow exponentially in terms of content size and hence must be supported with an efficient searching platform to facilitate fast lookup of the desired content. However, since these wikis contain highly sensitive matter, relying on external proprietary search engines such as Google, Bing is not possible. Companies, thus, rely heavily on existing open-sourced search engine platforms such as Lucene, Sphinx. Since the nature of the internal wikis can vary greatly, current user experience shows that the result produced by such search engine platform is often inaccurate. In this paper, we aim to present a search engine powered by ‘Learning to Rank’ system, having the capability to model its ranking algorithm according to the needs of the company.
KeywordsSearch engine Learning to Rank Hadoop Machine learning
- 1.He, C., et al.: A survey on learning to rank. In: International Conference on Machine Learning and Cybernetics, vol. 3. IEEE (2008)Google Scholar
- 4.Burges, C., et al.: Learning to rank using gradient descent. In: Proceedings of the 22nd International Conference on Machine learning. ACM (2005)Google Scholar
- 5.Boytsov, L., Belova, A.: Evaluating learning-to-rank methods in the web track adhoc task. In: TREC (2011)Google Scholar
- 6.Burges, C.J.C., Ragno, R., Le, Q.V.: Learning to rank with nonsmooth cost functions. In: NIPS, pp. 193–200 (2006)Google Scholar