en
0.25
0.5
0.75
1.25
1.5
1.75
2
Data Stream Mining for Ubiquitous Environments
Published on 2012-11-163455 Views
In the data stream computational model examples are processed once, using restricted computational resources and storage capabilities. The goal of data stream mining consists of learning a decision mo
Related categories
Presentation
Challenges in Ubiquitous Data Mining00:00
Outline02:13
Flags02:25
Motivation02:26
Problem Formulation: Network Data Model03:40
Querying Model04:11
Routing schemes04:45
The Demand for Learning05:02
Illustrative Example: Renewable Power Prediction05:54
Collaborative Forecasting: Main Idea06:59
Collaboration09:08
Local Search (1)10:25
Local Search (2)10:47
Local Search (3)10:50
Collaboration: broadcast time-stamps of similar contexts10:51
Local search: Inferring the Global Context (1)11:51
Local search: Inferring the Global Context (2)12:00
Local search: Inferring the Global Context (3)12:03
The Global Context12:08
Prediction12:16
Lessons Learned12:25
Clustering Distributed Data Streams (1)13:20
Clustering Distributed Data Streams (2)14:48
Clustering Distributed Sources of Data Streams15:17
Clustering Distributed Data Streams (3)15:44
Clustering Distributed Data Streams (4)16:27
System Overview16:42
Step 1: Local Step18:13
Local Adaptive Grid18:28
Step 2: Aggregation Step19:02
Monitoring States19:40
Frequent States22:55
Step 3: Centralized Cluster22:58
Furthest Point Clustering23:02
Illustrative Example24:36
Main Achievements25:00
A k-means Algorithm for Evolving Data25:24
Example: Local Clustering (1)26:35
Example: Local Clustering (2)27:35
Example: Local Clustering (3)27:48
Receiving Neighbors Data28:48
Sending Data to Neighbors29:07
After 512 Iterations...29:38
Evaluation: Electrical Grid Data30:02
A World in Movement31:22
The Challenges of UDM32:51
Thank you!33:37