A Case For Interaction: A Study of Interactive Information Retrieval Behavior and Effectiveness By Jürgen Koenemann and Nicholas J. Belkin (1996) John Clougherty1 Presented by John Clougherty
Experiment to Measure Effectiveness of Relevance Feedback and User Interaction John Clougherty2
Relevance Feedback and User Interaction Improves Retrieval Effectiveness Relevance feedback increased retrieval effectiveness Increased user interaction with relevance feedback made interactions more efficient John Clougherty3
Exponential Number of New Users are Gaining Access to IR Tools Information is everywhere and it’s availability is growing rapidly Anyone with a Computer/Smartphone/Tablet can use Bing to Google something New users have none or minimal training John Clougherty4
Increase in Computerized Tools to Support User’s Information Need User’s have an information need Search engines are designed to provide that information need Improvements in search engines are being developed every day Speed, Best-match, Relevance, Flexibility, Push and Pull John Clougherty5
Novice IR System Users Perform Query Reformulation INQUERY System Novice Users 2000 Documents Two Topics and Recall John Clougherty6
Four Levels of Information Retrieval System 2. Opaque Hidden Effect of Relevance Feedback No User Interaction John Clougherty7 4. Penetrable Visible Effect of Relevance Feedback User Interaction with Feedback 1. Baseline No Relevance Feedback 3. Transparent Visible Effect of Relevance Feedback No User Interaction
Penetrable System Performed Best for and Number of Iterations John Clougherty8
Relevance Feedback is Heavily Utilized but User Interaction is Minimal Directly or Indirectly obtained by the user Most relevance feedback algorithms are done implicitly User interaction has yet to be implemented in a big way John Clougherty9
Sources 33f1aec03e68.jpg Prologue-Hologram-2.jpg chart-300x300.jpg John Clougherty10
John Clougherty11 Thank You