Search result ranking method based on Borda algorithm
A technology of search results and sorting methods, applied in computing, special data processing applications, instruments, etc., can solve problems such as low precision rate, inaccurate sorting of search results, and inability to fully represent changes in correlation, and achieve precision rate Improve and reduce search time and network resource usage, calculate accurate results
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
specific Embodiment approach 1
[0034] The search result sorting method based on the Borda algorithm of the present embodiment combines figure 1 Shown flow chart, described method is realized through the following steps:
[0035] Step 1, define the set S={s of search engines of n members in the meta search engine 1 ,s 2 ,...,s n}; query term q; all search results r k The set R={r 1 ,r 2 ,...,r m}, the subscript k indicates the location, k=1,2,...,m, each search result r k Including URL, title, abstract and related scores, expressed as s in array form i _Url[k], s i _Title[k], s i _Abs[k] and s i _Score[k], i=1,2,...,n;
[0036]Step 2, in the search result list of the independent search engine obtained according to the correlation with the query word q, perform the search result r k Calculating the score of position k in the search engine, so that the position score is uniformly normalized;
[0037] Step 3. Combine the query word q with the search result r k The similarity weighted sum of the t...
specific Embodiment approach 2
[0040] The difference from Embodiment 1 is that in the search result sorting method based on the Borda algorithm in this embodiment, in step 2, in the search result list obtained by sorting according to the correlation with the query word q, perform the search result r k In the calculation of the score of position k in the search engine, the process of making the position score unified and normalized is as follows:
[0041] Each search result r in the search result list k The position k of the index greatly reflects the degree of relevance to the query word q, and the search result r in the search result list k The higher the position of the search results is, the higher the correlation with the query word q entered by the user is, so it is necessary to consider the position information of the independent search engine. In order to make the position score more accurate, the n search engine members S 1 , S 2 ,...,S n For query term q, search engine s j Return m search resul...
specific Embodiment approach 3
[0046] The difference from the specific embodiment 1 or 2 is that in the search result sorting method based on the Borda algorithm in this embodiment, the query word q and the search result r are combined in step 3 k The similarity weighted sum of the title and abstract of the search results r k The process of calculating the global similarity with the user input query word q is as follows:
[0047] Let the query term q have z feature items t 1 ,t 2 ,...,t z , and there is a document d 1 and document d 2 ; if in document d 1 In , a feature item appears many times, while other feature items do not appear, but in the document d 2 In , all z feature items appear once, although the document d 1 and document d 2 The word frequency is the same, but it is obvious that the document d 2 The most comprehensive coverage of feature items, document d 2 The situation is more relevant;
[0048] For example: for the query string "Central People's Government", the query word q is di...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More