then went to the last article, "station search two," to continue with the content of the structure and framework.
5, and then consider the arrangement of query results
, our station search completed the keyword analysis matching process, and now we want to sort out the good results. And the precise relevance ranking is the sign of full search. Most of the like query using SQL statement search is based on shelf time, according to sales, according to the price of a simple sort.
refers to the sort by relevance through the analysis of query relevance decision between keywords and text sorting, B2C search results sorting rules, should consider the following factors (brief introduction, far more than the:
1) how many query payment strings are included in the search results file?
2) how high do keywords occur in documents?
3) how close are keywords, such as: adjacent, same sentence, or same paragraph,
4) where keywords appear, and if they appear in headings, they are usually more relevant than they appear in the details of the goods.
5) query the popularity of the goods in which the keyword appears
to sort by relevance, be careful, users usually assume that the top few results are the best results, once the query is not satisfied, replace keywords. So: 1, should allow the user to choose the other; 2, ranking algorithm as the default on time, according to the order according to the sales.
6, sort simple here, to say is nonsense, we continue to consider in order to better experience, also have to consider what auxiliary function
common as follows:
1) spell check tool (more than chicken notebook)
2) pre input prompt
3) highlight query keyword
4) related queries recommended
7, continue to consider query results presentation
1) what content does the search result appear?
2) explains what you’ve been inquiring about,