Continuous Data Stream Processing

1
Continuous Data Stream Processing

• Make Lab

2
Application and Environment

• The framework and objectives
• Querying data streams
• Mining data streams
• An integrated system
• Music virtual channel
• V.C. player
• user profile (moving objects)
• user queries (continuous queries)
• Clustering engine
• virtual channel (collaborative decision)
• Filtering engine
• favorite channel (content-based decision)

Indexing server
Decoder server
Camera
UniGrid
Sensor
Video-on-demand
3
Research Directions (1/2)

• Mechanisms for processing continuous queries
• Temporal query processing
• Continuous query processing over event streams
  based on approximate matching mechanisms
• Continuously matching episodes for triggering
  episode rules over event streams
• Spatial query processing
• Continuous clustering moving objects in multiple
  space
• Monitoring heterogeneous kNN moving objects
  considering location-independent attributes
• Aggregate query processing
• An efficient method for processing multiple
  continuous Top-k queries
• Maintaining moving sums over data streams

4
Research Direction (2/2)

• One-pass mining algorithms for data streams
• Frequent tree pattern mining
• Discovering frequent tree patterns over data
  streams
• Mining frequent subtrees over data streams using
  closed subtrees
• Frequent Itemset mining
• Processing multiple queries of finding frequent
  itemsets over multiple data streams
• Mining frequent itemsets from data streams with a
  time-sensitive sliding window
• A novel hash-based approach for mining frequent
  itemsets over data streams with memory
  consideration under landmark model
• Frequent sequence mining
• Mining serial episode rules with successor lag
  times over multiple data streams

5
Continuous query processing over event streams
based on approximate matching mechanisms
UniGrid
UniGrid Portal
Filtering Engine
User
6
Processing multiple queries of finding frequent
itemsets over multiple data streams
7
Multiple queries of frequent itemsets over
multiple streams
UniGrid
A
UniGrid Portal
B
C
D
8
An efficient method for processing multiple
continuous Top-k queries
ltO1, 15gtltO2, 13gtltO3, 16gt
Multiple continuous top-1 queries ltO3, 16gt
9
Frequent tree patterns mining over data streams
Frequent Tree Patterns
STMer
T1