Computer Scientists Devise HITS Algorithm To Measure Human Pecking Order
@smriti-ZtAJsx
•
Oct 15, 2024
Oct 15, 2024
1.1K
Society has long associated power with money but measuring how influential a person is, is tad more complex than that. Algorithms have, to a great extent, helped make sorting easier. Google's PageRank is the quintessential application of one such algorithm, called the Hyper Induced Topic Search or HITS algorithm. HITS algorithm classifies important pages fall into two categories -hubs and authorities. A page is ranked higher if it points to other important pages and if other important pages point to that page.
This idea was introduced by John Kleinberg, a computer scientist at Cornell University in Ithaca and now he has come up with a new concept with his colleagues for measuring power and influence. This process is to be carried out by observing the style of language during a conversation. The key to this is linguistic co-ordination, in which speakers naturally copy the style of their interlocutors. Hence, in a group discussion, the one that responds in a word pattern similar to the interlocutor, is the one that gets influenced. This reveals the power differences between members of the group.
#-Link-Snipped-#
Functional words fall into categories such as articles, auxiliary verbs, conjunctions, high-frequency adverbs and so on. These are the words that distinguish the talking style of the speaker and responder. If the responder follows the same functional word pattern as speaker, then the speaker holds higher power in the conversation. Kleinberg and co analysed two types of text surrounding around an important debate: transcripts of oral arguments in the US Supreme Court and editorial discussions between Wikipedia editors. By noticing the changes in the linguistic styles, the team concluded that Admins become less likely to co-ordinate with others while lower ranking individuals are more likely to co-ordinate with admins.
Potential applications include ranking editors or a company using internal email records to determine which individuals display most power and influence.
Source & Image Credit: #-Link-Snipped-#
This idea was introduced by John Kleinberg, a computer scientist at Cornell University in Ithaca and now he has come up with a new concept with his colleagues for measuring power and influence. This process is to be carried out by observing the style of language during a conversation. The key to this is linguistic co-ordination, in which speakers naturally copy the style of their interlocutors. Hence, in a group discussion, the one that responds in a word pattern similar to the interlocutor, is the one that gets influenced. This reveals the power differences between members of the group.
#-Link-Snipped-#
Functional words fall into categories such as articles, auxiliary verbs, conjunctions, high-frequency adverbs and so on. These are the words that distinguish the talking style of the speaker and responder. If the responder follows the same functional word pattern as speaker, then the speaker holds higher power in the conversation. Kleinberg and co analysed two types of text surrounding around an important debate: transcripts of oral arguments in the US Supreme Court and editorial discussions between Wikipedia editors. By noticing the changes in the linguistic styles, the team concluded that Admins become less likely to co-ordinate with others while lower ranking individuals are more likely to co-ordinate with admins.
Potential applications include ranking editors or a company using internal email records to determine which individuals display most power and influence.
Source & Image Credit: #-Link-Snipped-#