Sparse Multitask Learning for Detecting Influential Nodes in
- Slides: 12
Sparse Multi-task Learning for Detecting Influential Nodes in an Implicit Diffusion Network Luca Lugini Publication by Yingze Wang, Guang Xiang, and Shi-Kuo Chang
Overview 1. Introduction 2. Linear Influence Model 3. Proposed Method: MSLIM 4. Experiments and Results 5. Conclusions
Introduction Context: social network analysis Goal of the paper: find the most influential nodes Why is this important?
Linear Influence Model Vk: Volume of topic k at time t Mk: Influence indicator matrix (set of nodes propagating topic k at time t) Iu: Influence function for node u
Linear Influence Model Influence function: Solution:
Limitations of LIM topic 1 1. Each topic is independent of each other 1. Single influence function for each node topic 2 topic k
Proposed Method: Multi-task Sparse Linear Influence Model (MSLIM) Extend LIM by introducing topic-sensitive influence functions Iu 1 t Iu t . . . Iu 2 t Iuk t
Proposed Method: Multi-task Sparse Linear Influence Model (MSLIM)
Proposed Method: Multi-task Sparse Linear Influence Model (MSLIM) Introducing regularization penalties topic-sensitive volume prediction penalty encouraging Iu to be zero penalty encouraging Iuk to be zero influential nodes for topic k
Experiments and Results Twitter dataset: 1000 users 3 years of tweets over 2. 6 million tweets 50 topics Results:
Conclusions Location distribution of influential nodes for topics “Apple” and “Samsung” Location distribution of influential nodes for topics “Startup” and “Finance”
Questions? References: - Wang, Yingze, Guang Xiang, and Shi-Kuo Chang. "Sparse Multi-Task Learning for Detecting Influential Nodes in an Implicit Diffusion Network. " AAAI. 2013. - “Learning with Sparsity for Detecting Influential Nodes in Implicit Information Diffusion Networks”, Yingze Wang, 2014.
- European physical characteristics
- Outliers and influential points
- Instrumental power examples
- Christian louboutin louis benech
- Sniffer for detecting lost mobiles
- Detecting evolutionary forces in language change
- Detecting golden ticket attack
- How do fraud symptoms help in detecting fraud
- Havex
- Cuadro comparativo e-learning b-learning m-learning
- Sparse conditional constant propagation
- Suatu array dikatakan lower triangular jika
- Sparse matrix operator kernel emissions