network science学习笔记——Barabási model和Bianconi-Barabási model

课程4基于Barabási的《network science》第5、6章。

之前的随机网络不能产生hubs的原因是N已固定(static models),而现实生活中节点总数一定是不断变化的,例如citation network。同时哪些点之间有link也完全是随机的,但是我们可以想见,当一个点的度越多时,其越有可能和其他点再连接(比如有影响力的论文会被越来越多地引用)。
基于这两点,Barabási model考虑growth(即节点总数的变化为增长)和preferential attachment来建立模型,越old的节点,其度随着t增长越大(度动力学)。且模型具有power law。度分布独立于t→度分布是稳定的;度分布独立于N(5.5 degree distribution)。如果缺了growth,那么度分布为稳定的指数分布,指数分布比幂函数衰减快得多,网络中很难出现hub;如果缺了preferential attachment,那么度分布没有稳定的分布,网络收敛成完全图。两个条件缺了任何一个都构成不了scale-free network。
总而言之,Barabási model生成了度指数γ=3的无标度网络。

(5.7对现实里的preferential attachment的度量,再好好理解一下)

然而!事情并没有那么简单,最开始的互联网搜索引擎并非google和facebook,但是现在它们一定是度相当大的节点。所以Barabási model里节点越old,度的值越大的说法显然是有问题的。因为会有结点”死亡“,会有结点突然特别受欢迎,所以需要另一个变量来刻画这个性质,即fitness。含这个特性的模型称为Bianconi-Barabási model。
当我们在growth步加入一个新的node时,不仅有m条新links,且有一个fitness值,这个值是不会变的。当我们在preferential attachment步,这个fitness值会和度的大小相乘,一起作为连接link的概率。
在Barabási model里,power exponent β=0.5,每个node的度按时间的平方根速度增长;
在Bianconi model里,node有各自的β,fitness高的node,degree增长快。时间够长,后出现但fitness大的node的度也会超过先出现但fitness小的node。

这两类模型刻画了网络随时间演化的机制,所以称为演化网络模型(evolving network models),前者是人们研究最多的模型,而后者更考虑到了节点更popular的特性-fitness适应性。

  • 1
    点赞
  • 1
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
"This is not just an important but an imperative project: to approach the problem of randomness and success using the state of the art scientific arsenal we have. Barabasi is the person." --Nassim Nicholas Taleb, author of the New York Times bestselling The Black Swan and Distinguished Professor of Risk Engineering at NYU An international bestseller In the bestselling tradition of Malcom Gladwell, James Gleick, and Nate Silver, prominent professor László Barabási gives us a trailblazing book that promises to transform the very foundations of how our success-obsessed society approaches their professional careers, life pursuits and long-term goals. Too often, accomplishment does not equal success. We did the work but didn't get the promotion; we played hard but weren't recognized; we had the idea but didn't get the credit. We convince ourselves that talent combined with a strong work ethic is the key to getting ahead, but also realize that combination often fails to yield results, without any deeper understanding as to why. Recognizing this striking disconnect, the author, along with a team of renowned researchers and some of the most advanced data-crunching systems on the planet, dedicated themselves to one goal: uncovering that ever-elusive link between performance and success. Now, based on years of academic research, The Formula finally unveils the groundbreaking discoveries of their pioneering study, not only highlighting the scientific and mathematic principles that underpin success, but also revolutionizing our understanding of: Why performance is necessary but not adequate Why "Experts" are often wrong How to assemble a creative team primed for success How to most effectively engage our networks And much more.

“相关推荐”对你有帮助么?

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值