Search Engine Evaluation Index-search engine technology

In traditional information retrieval, the basic index of the system: Recall (Recall) and precision ratio (pricision), recall is the number of related documents retrieved and the ratio of all relevant documents in the document library; The precision ratio is the percentage of the number of related documents retrieved and the total number of documents retrieved. For a retrieval system, recall rate and accuracy can not be the same: high recall rate, low precision, high precision, low recall.

For the search engine system, because no one search engine system can collect all the Web pages, so recall is difficult to calculate. Because of the characteristics of the Internet mass information, the results of query return are mostly very much, so the current search engine systems are very concerned about precision. At the same time, the information of the Internet is dynamic change, the search engine must reflect this change.
In short, search engine to meet the needs of users of information query, improve the user's search experience. Here are some of the more important indicators:
1. Web coverage. To improve the recall is the basis of ensuring precision.
2. The accuracy of the return results is mainly the accuracy of the first page results. Most users just look at the first page of the search results.
3. Filtering for repetitive information returns. The return result should be as much as possible without duplication and similar results.
4. Web page update speed. Depends on the discovery of new Web pages, and the Dead chain (refers to the inability to access the Web page) of the timely deletion. A large number of links to dead chains and outdated information in the results will degrade the user experience.
5. Search service corresponding time, that is, the user submitted to retrieve the results returned after the wait time, generally less than a second.
6. The system stability of the search service.