The linking partnering suggestion algorithm is a constantly evolving matching algorithm. The idea is to replicate some of the concepts used in major search engines. The basics for every search engine is that of indexing and ranking.
Indexing is the process of extracting keywords from a website. In poor words a website is read as an Html document. From this Html document one would remove the HTML syntax tags and common words like verbs and decorators. What is left is a set of words that appear in the web site. From this set of words an algorithm can determine what the site is about. In linking we are making use of a natural language processing service alchemy as thistakes the process of indexing to the next level. Thanks to this service we are able to determine the scoring of keywords based no only on number of occurrences but even on their context like for example if a keyword happens to be in a title tag then this is given more value. This concept is used by major search engines like Google search.
Once a website is published on line search engine crawlers will eventually index the site and based on several parameters like
Once the site is indexed and ranked and stored in our databases. The partnering algorithm will scan our internal database to make a match on the indexed keywords. From the top 3 matching sites if these are above the threshold, that means that for our algorithm they are close enough to contain similar topics then they will be considered as a potential link suggestion. The system will check if the site that is likely to be suggested has already linking history and if it accepted any linking suggestions. This is done so that web masters are encouraged to give feedback on the link suggestions proposed. Also the link suggestion algorithm takes into consideration the ranking values of the site so as match sites with similar scoring.