IDS 200 Lecture Notes - Lecture 20: Knowledge Graph, Common Interface, Database Caching

• School: UIC
• Department: Information &Decision Sciences
• Course: IDS 200
• Professor: Lundquist, Doug E.

Document Summary

At a basic level, pages can be indexed just by search terms. Works for very rare terms on mostly-text pages. Not for popular searches on object-heavy pages. Remember: google is working semantically to match the user"s intent. Indexing can incorporate semantics, by creating entries for term groups or concepts. For example, buy car might be empirically observed to mean a search for car dealerships rather than for how to or price comparisons. These may be integrated into normal search or displayed separately via the knowledge graph. Search engines consider terms" usage and occurrence. For example, as humans we know that charles babbage or. A book index would store babbage, charles , not separate entries for charles and babbage . As readers, we would see the difference between or. Index entries are semantic units , one or more words with a specific meaning. Determining the intended or most likely grouping helps.