By Yizhou Sun,Jiawei Han
In this booklet, we examine the rules and methodologies of mining heterogeneous info networks. Departing from many present community versions that view interconnected info as homogeneous graphs or networks, our semi-structured heterogeneous info community version leverages the wealthy semantics of typed nodes and hyperlinks in a community and uncovers unusually wealthy wisdom from the community. This semi-structured heterogeneous community modeling ends up in a chain of recent ideas and robust methodologies for mining interconnected information, together with: (1) rank-based clustering and class; (2) meta-path-based similarity seek and mining; (3) relation strength-aware mining, and plenty of different strength advancements. This e-book introduces this new learn frontier and issues out a few promising study directions.
Table of Contents: creation / Ranking-Based Clustering / type of Heterogeneous info Networks / Meta-Path-Based Similarity seek / Meta-Path-Based dating Prediction / Relation Strength-Aware Clustering with Incomplete Attributes / User-Guided Clustering through Meta-Path choice / examine Frontiers
Read or Download Mining Heterogeneous Information Networks: Principles and Methodologies PDF
Similar data mining books
This specialist compilation offers a set of winning database advertising methodologies for large info. It deals suggestions to universal difficulties within the database advertising and marketing undefined, concentrating on the wishes of knowledge analysts and knowledge miners. The quantitative options defined marry conventional statistical methodologies with new computer studying tools.
Grasp Oracle company Intelligence 11g stories and Dashboards bring significant company details to clients every time, wherever, on any gadget, utilizing Oracle enterprise Intelligence 11g. Written via Oracle ACE Director Mark Rittman, Oracle company Intelligence 11g builders advisor absolutely covers the newest BI record layout and distribution options.
Even if clients tend to settle for the suggestions supplied via a recommender process is of maximum value to procedure designers and the dealers who enforce them. by way of conceptualizing the recommendation looking and giving dating as a essentially social method, very important avenues for realizing the persuasiveness of recommender platforms open up.
This paintings offers an cutting edge examine using open facts for extracting details to observe and stop crime, and likewise explores the hyperlink among terrorism and arranged crime. In counter-terrorism and other kinds of crime prevention, foresight approximately capability threats is extremely important and this knowledge is more and more on hand through digital facts resources akin to social media communications.
Extra info for Mining Heterogeneous Information Networks: Principles and Methodologies
Mining Heterogeneous Information Networks: Principles and Methodologies by Yizhou Sun,Jiawei Han