The word social is currently very popular, we have social networks, social software, social bookmarking, and now social algorithms. Social networking is an activity, social software is a tool that makes social networking possible, but what is a social algorithm?
The following tests are used to define the meaning of social algorithms.
Social network.
Let's take a look at these examples. Notable social networking sites include Flickr, YouTube, MySpace, Friendster, Wikipedia, Facebook, del.icio.us and Second Life's virtual gaming world. Bookmarks and tags are considered social network activities. The purpose of the web is to let people gather, collaborate or share something [photos, music, movies, information, etc.].
social application.
Social software is a tool that enables people to connect to the Internet. These include email, mailing lists, RSS, IRC, instant messaging, Napster, such as P2P, blogs, wikis, AJAX, Web 2, etc. They are characterized by the network must be interactive, bottom-up [user-provided content]. These software were formerly called components. Some of these softwares are old, but others are like blog updates. The wiki has been around since the nick's father, Ward Cunningham, has been used, but it is only considered a geek and has only been widely accepted in recent years. Some software uses wiki records, and businesses, educational institutions have begun to use it.
algorithm.
So what is a social algorithm? The algorithm is similar to a cooking recipe or computer program with step-by-step instructions for executing the process. The algorithms are listed in pseudocode, are easy to understand, and are more abstract than computer programs. These programs are said to implement some algorithms, namely machine-level translation of pseudocode.
Although most algorithms are digital, they are not necessarily shown as cooking recipes, logical unification algorithms, string matching, face recognition, and so on.
Based on the agent.
Social algorithms differ from general algorithms in that they involve agents and algorithms are the result of agent interactions. The ant colony algorithm is an example of an ant agent that solves problems such as shortest path or traveling salesman problems. Social algorithms can be used as distributed problem solving as ant colony algorithms, but not necessarily.
It is said that Google's PageRank algorithm is the decisive factor for Google to win online battles. The PageRank algorithm uses a number of criteria to rank websites, including the number of inbound links. The weight of each portal link is the PageRank of the referenced site [the algorithm is recursive, almost real-time].
To understand the algorithm, here is a description:
Lawrence Page and Sergey Brin describe the original PageRank algorithm in several publications.
It consists of. Given
PR[A]=[1-d]+ d[PR[T1]/ C[T1]+ ... + PR[Tn]/ C[Tn]
Where?
from
PR[A] is the PageRank of page A,
from
PR[Ti] is the PageRank of the page Ti linked to page A,
from
C[Ti] is the number of outbound links for page Ti and
from
d is the damping coefficient and can be set between 0 and 1.
Since PageRank defines the ranking of a website in search, it affects a lot of things, and in the final analysis, many people get revenue from the website. Here, people acting as agents try to modify their site parameters to increase their ranking. This is often referred to as SEO: Search Engine Optimization.
This is an example of an algorithm that provides rules for a social network, and doing so modifies proxy behavior or actions. There has been a case of abuse of the algorithm, in one case, someone built a website based on [almost] links, only without content, but managed to get a high PageRank. Since the PageRank algorithm is also evolving, this anomaly may be resolved.
Internet auctions and reverse auctions provided by EBay are also algorithms that provide game rules, and we are players.
Digg.com is a social bookmarking site, the initial suggestion is 1 digg, if they like this suggestion, people can dig again, if they don't like it, you can dig it. The number of diggs indicating the popularity of the recommendation. The system will help the collaborating people to mine each other's recommendations.
We see that social algorithms have weaknesses that are often used by some people. This is the need to improve social algorithms.
Netflix is a well-known case of providing $1 million to anyone who can increase the accuracy of existing algorithms by 10%.
Non-human players and avatars.
Back to the definition of social algorithms as a multi-agent based algorithm in which the agent is a human, we believe that the definition must be expanded to include situations where the person is replaced by an animal or software substitute. I would consider using the Digg algorithm for simulation, or using smart agents as a market trading simulation for social algorithms. The same is true of the ant colony algorithm, which replaces animals with humans. The same is the group, the flock, etc.
Cellular automata is a boundary case that is agent-based but is often used to simulate physical and chemical processes. If used to mimic human activity, it would be a social algorithm.
The evolutionary algorithm that is abstracted as a Darwin process is also a social algorithm.
from
Genetic algorithms and their variants are included here.
Evolutionary game theory.
The field of social algorithms intersects with evolutionary game theory. Game theory studies strategies used by agents, such as trading, auctions, marketing, voting, etc. The prisoner's dilemma set a game for two prisoners, but it can be iterated and played in a crowd and then become a social algorithm.
If you've ever played Second Life, you know how complex social algorithms are. In the virtual world, the entire life, including the economy, relationships, and property, is defined by algorithms.
Orignal From: Defining social algorithms
No comments:
Post a Comment