Kno wledge Disco v ery in Spatial Data b y Means of ILP - - PowerPoint PPT Presentation

▶
kno wledge disco v ery in spatial data b y means of ilp
SMART_READER_LITE
LIVE PREVIEW

Kno wledge Disco v ery in Spatial Data b y Means of ILP - - PowerPoint PPT Presentation

Apr 01, 2023 388 likes •490 views

Kno wledge Disco v ery in Spatial Data b y Means of ILP Lub o Pop elnsk Masaryk University Brno and CTU Pr ague, Cze chia Email: popel@fi.muni.cz Motiv ation Inductiv e logic programm ing and

slide-1
SLIDE 1 Kno wledge Disco v ery in Spatial Data b y Means
  • f
ILP Lub
  • Pop
elnsk Masaryk University Brno and CTU Pr ague, Cze chia Email: popel@fi.muni.cz Motiv ation
  • Inductiv
e logic programm ing and inductiv e query languages
  • Description
  • f
(ma yb e) inexactly dened geographic
  • b
jects
slide-2
SLIDE 2 Kno wledge Disco v ery in Spatial Data b y Means
  • f
ILP Lub
  • Pop
elnsk Masaryk University Brno and CTU Pr ague, Cze chia Email: popel@fi.muni.cz Outline 1. Inductiv e query language 2. Metho d & WiM 3. Examples 4. Discussion & F uture researc h
slide-3
SLIDE 3

Object1 Object 2 HIGHWAY_BRIDGE LINEAR BRIDGE Geometry Importance ROAD RIVER RAILWAY PLANAR Geometry FORESTRY BUILDING FOREST WOOD FOREST_HOUSE Forest Named State Named

Ob ject-orien ted database sc hema
slide-4
SLIDE 4 Ra w data 4
slide-5
SLIDE 5
  • b
ject descriptions (in F-logic) class descriptions (in F-logic) inductiv e query ? TRANSLA TE
  • @
@ R Example set Bac kground kno wledge T yp e denitions @ @ R
  • WiM
? result
  • f
inductiv e query (Horn clauses) GW iM sc hema 5
slide-6
SLIDE 6 WiM inductiv e learner ecien t searc hing for the renemen t graf shift
  • f
syntactic bias generator
  • f
near-misses
  • racles
needs from 2 to 4 examples for most
  • f
the ILP b enc hmark predicates (list pro cessing) smaller dep endency
  • n
the qualit y
  • f
the example set in comparison to some
  • f
ILP programs has b een tested b
  • th
  • n
go
  • d
examples and
  • n
randomly c hosen example sets. 6
slide-7
SLIDE 7 Inductiv e language extract < KindOfRule > rule for < NameOfT arget > from [ < ListOfClasses >] [< Constrain ts >] [ from p
  • in
t
  • f
view < Domain >] extract characteristic rule
  • extract
discriminate rule
  • extract
dependency rule
  • 7
slide-8
SLIDE 8 Discrimination
  • f
F
  • rests
and W
  • ds
Find a dierence b et w een forests and w
  • ds
from the p
  • in
t
  • f
view
  • f
area. ar ea is the name
  • f
set
  • f
predicates lik e ar ea(Geometr y ; Ar ea). extract discriminate rule for isF
  • rest
from forest in con trast to w
  • d
from p
  • in
t
  • f
view area. forest(F) :- geometry(F,GForest), area(GForest,Area), 100 < Area. Relation b et w een F
  • rests
and W
  • ds
Find a relation b et w een forests and w
  • ds
from the p
  • in
t
  • f
view
  • f
area. ar ea is the name
  • f
set
  • f
predicates lik e ar ea(Geometr y ; Ar ea). extract dep endency rule for forestOrW
  • d
from forest, w
  • d
from p
  • in
t
  • f
view area. forestOrWood(F,W) :- geometry(F,GF),area(G F,F A), geometry(W,GW), area(GW,WA), WA<GA. 8
slide-9
SLIDE 9 Discussion 1. The query language is quite p
  • w
erful ) quite complex queries can b e form ulated. Ho w ev er, the price that user has to pa y for is sometim es to
  • big.
2. Ho w to pro cess large amoun t
  • f
data F uture researc h
  • In
terface to P
  • stgreSQL
  • b
ject-relational DBMS
  • Geographic
domain kno wledge 9