Web Search based on Ranking

bg-new
Author/s

Ioannis Pitas (AUTH)

About the resource/s

This lecture overviews Web Search based on Ranking that has many applications in Web Science and Social Media Analytics. It covers the following topics in detail: Architecture of Web Search Engine: Crawler, Indexer, Query Processor. Timeline of Ranking at Indexed Pages. Ranking Algorithms: Based on Frequency (TF-IDF), Based on Graph-Link Analysis (PageRank, Hits, SalsaUsersRank, SimRank). Neural Networks-based ranking algorithms: Pointwise (Ranking with Large Margin Principles-SVM), Pairwise (SortNet)Listwise (ListNet). Current Neural Networks for Ranking: GEPSMatchPyramid, C-DSSM, NeuBase. Deep Learning MetricsAngular Loss, Nearest Neighbors Gaussian Kernels.

Other Sources