bigdata large-scale data mining machine learning/ …...graph mining is to discover hidden...

1
Copyright © 2013 NTT. All Rights Reserved. <Contact>[email protected] S47 Large-scale data mining Efficiently Discover Hidden “Relationships” in Big Data BigData Machine learning/ Data analysis Our graph mining algorithms achieve the fastest speed in the world (more than 10 times faster than existing algorithms). The algorithms handle large-scale graphs with hundreds of millions of nodes with a single PC server. The supported graph mining algorithms are clustering, Personalized PageRank, and graph diameter analysis. The algorithms are implemented as a library of two general- purpose programming languages; Java and JavaScript. The library is available as a plug-in in graph analysis tool “Gephi”. Applicable for community discovery, user or content recommendation in social graphs, such as SNS and twitter, and in web page search. Useful in general graph analysis made by Gephi. Features Application Scenarios NTT Group Global Advantage NTT has developed graph mining algorithms achieving the fastest speed in the world. Improving the speed of graph mining makes it possible to be applied for large-scale data, so called big data. We have developed efficient graph mining algorithms for analyzing large-scale graph data such as social graphs. Graph mining is to discover hidden relationships in graphs by analyzing the structure of the graphs. We achieve high-speed graph mining for large-scale data such as clustering, Personalized PageRank, and graph diameter analysis. Clustering Graph mining library Personalized PageRank Metrics (Graph diameter, etc) Various graph data Social graph (Facebook, Twitter) Web pages Applications API - Java - JavaScript Efficient algorithms - Graph analysis made by Gephi - Community discovery and recommendations Implemented in general-purpose programming languages Discover closely connected groups Discover important nodes Estimate width of graphs by calculating the maximum number of hops Node Edge Baseball community Influencer Pair of persons, they communicate most infrequently Soccer community Social graph Social graph Social graph

Upload: others

Post on 05-Jul-2020

0 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: BigData Large-scale data mining Machine learning/ …...Graph mining is to discover hidden relationships in graphs by analyzing the structure of the graphs. We achieve high-speed graph

Copyright © 2013 NTT. All Rights Reserved.

<Contact>[email protected]

S-47

Large-scale data mining

Efficiently Discover Hidden “Relationships” in Big Data

BigData

Machine learning/ Data analysis

■ Our graph mining algorithms achieve the fastest speed in the world (more than 10 times faster than existing algorithms).

■ The algorithms handle large-scale graphs with hundreds of millions of nodes with a single PC server.

■ The supported graph mining algorithms are clustering, Personalized PageRank, and graph diameter analysis.

■ The algorithms are implemented as a library of two general-purpose programming languages; Java and JavaScript.

■ The library is available as a plug-in in graph analysis tool “Gephi”.

■ Applicable for community discovery, user or content recommendation in social graphs, such as SNS and twitter, and in web page search.

■ Useful in general graph analysis made by Gephi.

Features

Application Scenarios

NTT Group Global Advantage NTT has developed graph mining algorithms achieving the fastest speed in the world. Improving the speed of graph mining makes it possible to be applied for large-scale data, so called big data.

We have developed efficient graph mining algorithms for analyzing large-scale graph data such as social graphs. Graph mining is to discover hidden relationships in graphs by analyzing the structure of the graphs. We achieve high-speed graph mining for large-scale data such as clustering, Personalized PageRank, and graph diameter analysis.

Clustering

Graph mining library

Personalized PageRank

Metrics (Graph diameter, etc)

Various graph data

Social graph (Facebook, Twitter)

Web pages

Applications

API - Java - JavaScript

Efficient algorithms

- Graph analysis made by Gephi - Community discovery and recommendations

Implemented in general-purpose programming languages

Discover closely connected groups

Discover important nodes

Estimate width of graphs by calculating the maximum number of hops

Node

Edge

Baseball community

Influencer

Pair of persons, they communicate most

infrequently

Soccer community

Social graph

Social graph

Social graph