More Related Content
Similar to Evaluation of search engine (20)
More from Dr. B T Sampath Kumar (20)
Evaluation of search engine
- 1. Dr B T Sampath Kumar Professor Department of Library and Information Science Tumkur University, Tumakuru, INDIA www.sampathkumar.info Evaluation of Search Engines
- 2. What is evaluation? • It is a process of assessing the quality of any thing. • Making judgment about the value of something. • It is a systematic determination of a subject's merit, worth and significance, using criteria governed by a set of standards.
- 3. Criteria for evaluation of Search engine • Evaluation involves assessing the performance of search engines to improve their effectiveness. • Some of the criteria for the evaluation of search engines are: – Recall – Precision – Novelty/Up-to-Datedness – Searching time
- 4. Cont.. – Database – Relevancy – Indexing – User Interface – Query functionality – Help menu and advanced search option.
- 5. Recall • It is an ability of a search engine to retrieve all relevant search results from its database. • Recall is also defined as the fraction of relevant pages that are retrieved. Database (relevant and irrelevant) Relevant pages Recall
- 6. What should be the recall of a search Engine? • Low? or • High? Database (relevant and irrelevant) More number of Relevant pages High Recall Less number of relevant pages Low Recall
- 7. Recall ratio Recall= 990 X 100 =73.33 1350 Recall ratio Number of relevant pages retrieved Total Number of relevant pages in the database 100
- 8. Precision • It is an ability of a search engine to withhold (hide) irrelevant search results. • Precision is the fraction of retrieved pages that are relevant.
- 9. Precision ratio Precision ratio= 800 X 100 =80 1000 Precision ratio Number of relevant pages retrieved Total Number pages retrieved 100
- 10. Novelty/Up-to-datedness • Up-to-datedness is an important criteria for revaluating a search engine. • The search engine needs to keep its index fresh. • Secondly, up-to-datedness factors are used in the ranking of Web pages.
- 11. Searching time • It is a time taken by a search engine to give relevant results to a search query.
- 12. Database • Database of a search engine is very important component. • The data from disparate sources plays well together for completeness and relevancy during the search process.
- 13. Relevancy • Relevancy ranking is the process of sorting the document results so that those documents which are most likely to be relevant to your query are shown at the top. • Relevancy depends on consistently testing and improving algorithms. • The better your understanding of the user intent, the higher search relevancy you can reach with your search engine.
- 14. Indexing • Crawls of the search engines are essential for indexing the content. • Some aspects of an indexing task that you should be aware of and evaluate include: –Speed of indexing –Indexing latency –Dynamic fields
- 15. User interface • The user interface configuration is just as critical as the back-end configuration. • Search engines need to have a user-centric, intuitive interface with which users are familiar and able to conduct search and analysis productively.
- 16. Query functionality • The search engine should able to support and optimize query-based search functions, depending on the types of data.
- 17. Help menu and advanced search option • The search engine must have help menu and advanced search options to search information available on the web.