To investigate the automatic construction of the high degree of correlation within the chain method

in the chain I say everyone should know. Within the chain (the chain too) good correlation is very important index. Based on the discuz program as an example, on how to use coreseek full text search in chain system construction of high correlation.

the first to solve the. We can use the target with the title of the article title or the full-text database of existing use like Levenshtein algorithm for comparison, find out the high correlation before one or several to do within the chain, rather than simply capture a specific keywords or tag.

Without the importance of

three, the discuz as an example to explain how to achieve.

?How to determine the level of correlation problem

, a status of

is now in the automatic production chain there are two kinds of methods: one is the chain by inserting the same tag; two is the pre specified keywords, automatic insertion method within the chain to do within the chain meet relevant keywords in the article.

1, first of all to coreseek before the correct installation, installation method can find the < reference website or my website four clothing network www#4ji#cn; < set up under Discuz coreseek full text search server and personalized self participle thesaurus > > the post. The installation looks very complex, it is not difficult to install, follow the tutorial step by step, do it.

two, how can

but these two methods have disadvantages, the former method is the automatic generation of tag if it is often used to generate a large number of no tag related degree is not guaranteed, if you manually set the tag is a very tedious work for a long time; the latter method to manually preset key words, a tedious secondly, there is no default key is not automatically generated within the chain, the method of three

I understand coreseek full-text search system uses the phrase similarity and Statistics (BM25) based on the composite Ranking method can easily and efficiently achieve this function, we can easily use it, can also have the ability to write their own procedures to achieve.

2, write a d>

is to meet the relevant keywords to generate the same chain correlation is not high.

disuz comes with the related posts are divided into two types. An implementation of tag, has said above shortcomings; another aspect is to search, but cheating is the related posts is implemented in JavaScript, the Shanghai dragon basically useless.

idea is as follows: when post is accessed automatically with the title of the post to coreseek search, find the matching degree of the article. Then to find the article add a link to the post.

In the chain of high correlation and automatic generation of

Leave a Reply

Your email address will not be published. Required fields are marked *