Compositions of Technologies

Having surveyed the major techniques, I turn now to questions regarding their synthesis. In particular, the questions that need to be asked are:

• What performance gain is there in combining term weighting and term selection techniques under the same model?
• Can term weighting and term selection techniques devised under different models be combined effectively?
• In constructing a system from scratch that would use Relevance Feedback, what would be the best way to build it?
• Rather than choosing the ``best'' weighting or selection technique, can better results be obtained by merging the output of multiple redundant techniques?
• What are the common assumptions from all these techniques? Do the answers for the above questions still hold if the underlying assumptions are relaxed?

 
• Combining Weighting and Selection under Similar Models
• Combining Weighting and Selection under Different Models
• Best Methods
• Combining Redundant Techniques
• Unstated Assumptions
  • Documents and abstracts are ``well-formed''
  • Users queries are ``well-formed''
  • Users make ``perfect'' choices
  • Closed World Assumption