Influence et cont...
Follow
Find tag "pagerank"
4.7K views | +2 today
Influence et contagion
L'influence et la contagion dans la cyberculture
Curated by luiy
Your new post is loading...
Your new post is loading...
Scooped by luiy
Scoop.it!

#Google matrix analysis of directed networks | #datascience #algorithms

#Google matrix analysis of directed networks | #datascience #algorithms | Influence et contagion | Scoop.it
luiy's insight:

This review describes matrix tools and algorithms which facilitate classification and information retrieval from large networks recently created by human activity. The Google matrix formed by links of the network has typically a huge size. Thus, the analysis of its spectral properties including complex eigenvalues and eigenvec- tors represents a challenge for analytical and numerical methods. It is rather surprising, but the class of such matrices, belonging to the class of Markov chains and Perron-Frobenius operators, was practically not inves- tigated in physics. Indeed, usually the physical prob- lems belong to the class of Hermitian or unitary ma- trices. Their properties had been actively studied in the frame of Random Matrix Theory (RMT) (Akemann et al., 2011; Guhr et al., 1998; Mehta, 2004) and quantum chaos (Haake, 2010). The analytical and numerical tools developed in these research fields allowed to understand many universal and peculiar features of such matrices in the limit of large matrix size corresponding to many-body quantum systems (Guhr et al., 1998), quantum comput- ers (Shepelyansky , 2001) and a semiclassical limit of large quantum numbers in the regime of quantum chaos (Haake, 2010). In contrast to the Hermitian problem, the Google matrices of directed networks have complex eigenvalues. 

more...
No comment yet.
Scooped by luiy
Scoop.it!

Onalytica: #BigData Influencers | #SNA #pagerank

Onalytica: #BigData Influencers | #SNA #pagerank | Influence et contagion | Scoop.it
luiy's insight:

The latest Onalytica ranking of Big Data influencers on Twitter is now available. The top 200 influencers are listed below.

 

Methodology: We took the tweets for the last 6 months containing the hashtag #BigData and treated mentions of other tweeters as a link to these. This initial network was made up of 150,486 nodes. After removing smaller, isolated components of the network we calculated mathematical PageRank, which is used to rank the list.

 

The methodology is different from our previous Big Data ranking which was based on Betweenness Centrality. This methodology weights connectors higher, but we have found that PageRank more effectively highlight the real influencers in the network.

 

The new entries in the top 200 are more due to people moving from below the 200 water mark to above because they are drawing more attention in the context of Big Data, than due to changes in the methodology.

more...
No comment yet.