en
0.25
0.5
0.75
1.25
1.5
1.75
2
From Inductive Querying to Declarative Modeling for Data Mining
Published on Jan 16, 20132774 Views
In this talk I shall present a personal perspective on the quest for a unifying framework and theory of data mining. The starting point will be the notion of an inductive database as proposed in the s
Related categories
Chapter list
From Inductive Databases to Declarative Modeling for Data Mining00:00
Data mining ...00:44
And do it again ...01:22
Problems01:28
Two frameworks02:27
Inductive Databases04:02
Data mining ...04:06
Inductive databases04:49
The vision06:08
A long-term perspective07:32
The Inductive Database framework08:43
Inductive database abstraction09:13
Abstraction09:50
Mine Rule Example (1)10:32
Mine Rule Example (2)11:14
Mine Rule Example (3)12:25
Mine Rule Example (4)13:06
DMQL Han & Kamber 2001 (m-k)13:41
Mining views (1)14:36
Mining views (2)16:34
A modern inductive database17:39
How do people get into town? What is the map of their trips in space and time? (1)18:00
How do people get into town? What is the map of their trips in space and time? (2)19:15
How do people get into town? What is the map of their trips in space and time? (3)19:41
A modern inductive query language19:44
Inductive databases - perspective20:04
Declarative Modeling22:33
The Challenge22:42
Key Claim23:11
The What, Why and How of Declarative Modeling23:50
What is declarative modeling ?24:00
How does it work ?25:30
Why declarative modeling ?26:20
Constraint Programming27:37
Constraint Satisfaction28:20
Solvers for CP28:58
Constraint Programming30:55
What about ML/DM ?31:38
Observation 1 (1)31:45
Data Mining32:06
Observation 1 (2)32:40
Observation 2 (1)32:58
Observation 2 (2)33:51
Observation 334:32
Evidence The case of Pattern mining35:21
Pattern Mining (1)35:38
Pattern Mining (2)36:50
A. Frequent Pattern Mining37:37
A. Frequent Itemset Mining (1)37:38
A. Frequent Itemset Mining (2)37:40
Frequent Itemset Mining39:44
Closed Itemset Mining40:19
Further Constraints41:05
How does it work ?41:22
Solver 141:38
Encoding in Zinc42:01
Resulting Search Strategy akin to Zaki’s Eclat [KDD 97]42:42
Solver 243:26
B. Correlated Pattern Mining44:22
Top-k Correlated Pattern Mining Subgroup Discovery44:27
Modeling perspective44:38
C. Pattern Set Mining45:30
Pattern Sets (1)46:19
Pattern Sets (2)46:22
Pattern Sets (3)46:48
k-Pattern Set Mining (|P|=k)47:28
Generality48:09
Pattern Mining48:30
http://dtai.cs.kuleuven.be/CP4IM49:06
Perspective49:19
All this is fine but...49:21
Efficiency / Scalability49:53
Task / Representation51:45
kLOG [Frasconi, Costa, DR, De Grave 12]52:26
E/R-MODEL53:07
Graphicalization53:25
kLOG (1)54:10
kLOG (2)54:12
kLOG (3)54:47
What if we succeed ?55:18
Our work today ...55:22
Our work tomorrow ... (1)55:39
Our work tomorrow ... (2)56:50
Conclusions (1)57:50
Conclusions (2)58:18
Questions ?58:19