Enterprise Search
• Search within documents of an organization, agency or company
• The search is done in the documents that exist in the intranet or on the internet.
• Takes document versions into account.
• Good performance in low hyperlink structure.
• Covers specific document formats per site or organization.
• Not based only on the document quality.
• Fewer users.
• Fewer relevant documents
• Deals with document restrictions
• Little redundancy
• It is important to have a human/admin rule in the process. It also requires domain experts who do query boosting (best bets)
• It needs taxonomies or metadata.
• Deals with all kinds of data
• Do not frequently change the data structure
• Confidentiality is very important
• The recall is more important for evaluation
• No spam
Web Search
• The search is done in the world-wide-web.
• Fails usually to satisfy site search requirements
• Poor in dealing with document versions
• Lower performance in low hyperlink structures
• Doesn’t cover all the document formats in organizations and sites.
• The quality is usually based on the document itself.
• A huge number of users.
• Higher redundancy
• Deals mainly with unstructured and semi-structured data.
• Changes the data structure frequently
• Precision is more important for evaluation
• A lot of spam