Difference between revisions of "Orange: CN2 Rule Viewer"

From OnnoWiki
Jump to navigation Jump to search
 
(4 intermediate revisions by the same user not shown)
Line 1: Line 1:
 
Sumber: https://docs.biolab.si//3/visual-programming/widgets/visualize/cn2ruleviewer.html
 
Sumber: https://docs.biolab.si//3/visual-programming/widgets/visualize/cn2ruleviewer.html
 
  
  
 
CN2 Rule Viewer
 
CN2 Rule Viewer
  
Inputs
+
==Input==
  
    Data: dataset to filter
+
Data: dataset to filter
    CN2 Rule Classifier: CN2 Rule Classifier, including a list of induced rules
+
CN2 Rule Classifier: CN2 Rule Classifier, including a list of induced rules
  
Outputs
+
==Output==
  
    Filtered Data: data instances covered by all selected rules
+
Filtered Data: data instances covered by all selected rules
  
A widget that displays CN2 classification rules. If data is also connected, upon rule selection, one can analyze which instances abide to the conditions.
+
Widget CN2 Rule Viewer dapat menampilkan CN2 classification rules. Jika ada data yang tersambung, kita bisa menganalisa instance mana yang patuh pada kondisi yang di atur oleh rules.
  
[[File:CN2RuleViewer-stamped.png|center|200px|thumb]]
+
[[File:CN2RuleViewer-stamped.png|center|600px|thumb]]
  
    Original order of induced rules can be restored.
+
* Original order of induced rules can be restored.
 
+
* When rules are many and complex, the view can appear packed. For this reason, compact view was implemented, which allows a flat presentation and a cleaner inspection of rules.
    When rules are many and complex, the view can appear packed. For this reason, compact view was implemented, which allows a flat presentation and a cleaner inspection of rules.
+
* Click Report to bring up a detailed description of the rule induction algorithm and its parameters, the data domain, and induced rules.
    Click Report to bring up a detailed description of the rule induction algorithm and its parameters, the data domain, and induced rules.
 
  
 
Additionally, upon selection, rules can be copied to clipboard by pressing the default system shortcut (ctrl+C, cmd+C).
 
Additionally, upon selection, rules can be copied to clipboard by pressing the default system shortcut (ctrl+C, cmd+C).
Line 27: Line 25:
 
==Contoh==
 
==Contoh==
  
In the schema below, the most common use of the widget is presented. First, the data is read and a CN2 rule classifier is trained. We are using titanic dataset for the rule construction. The rules are then viewed using the Rule Viewer. To explore different CN2 algorithms and understand how adjusting parameters influences the learning process, Rule Viewer should be kept open and in sight, while setting the CN2 learning algorithm (the presentation will be updated promptly).
+
Workflow di bawah adalah penggunaan sederhana widgetCN2 Rule Viewer. Pertama-tama, menggunakan widget File, data di load dan CN2 rule classifier di train. Kita menggunakan dataset titanic untuk membangun rule. Rules dapat dilihat menggunakan widget Rule Viewer. Untuk mengexplorasi berbagai alogitma CN2 dan mengerti bagaimana pengaturan parameter mempengaruhi proses learning, widget Rule Viewer harus terus open dan dalam penglihatan. Sehingga saat setting CN2 learning algorithm (hasilnya dapat langsung terupdate).
 
 
[[File:CN2-Viewer-Example1.png|center|200px|thumb]]
 
 
 
Selecting a rule outputs filtered data instances. These can be viewed in a Data Table.
 
 
 
  
 +
[[File:CN2-Viewer-Example1.png|center|600px|thumb]]
  
 +
Pemilihan sebuah rule akan mengeluarkan data instance tertentu yang terfilter. Ini dapat dilihat di widget Data Table.
  
 
==Referensi==
 
==Referensi==

Latest revision as of 09:38, 9 April 2020

Sumber: https://docs.biolab.si//3/visual-programming/widgets/visualize/cn2ruleviewer.html


CN2 Rule Viewer

Input

Data: dataset to filter
CN2 Rule Classifier: CN2 Rule Classifier, including a list of induced rules

Output

Filtered Data: data instances covered by all selected rules

Widget CN2 Rule Viewer dapat menampilkan CN2 classification rules. Jika ada data yang tersambung, kita bisa menganalisa instance mana yang patuh pada kondisi yang di atur oleh rules.

CN2RuleViewer-stamped.png
  • Original order of induced rules can be restored.
  • When rules are many and complex, the view can appear packed. For this reason, compact view was implemented, which allows a flat presentation and a cleaner inspection of rules.
  • Click Report to bring up a detailed description of the rule induction algorithm and its parameters, the data domain, and induced rules.

Additionally, upon selection, rules can be copied to clipboard by pressing the default system shortcut (ctrl+C, cmd+C).

Contoh

Workflow di bawah adalah penggunaan sederhana widgetCN2 Rule Viewer. Pertama-tama, menggunakan widget File, data di load dan CN2 rule classifier di train. Kita menggunakan dataset titanic untuk membangun rule. Rules dapat dilihat menggunakan widget Rule Viewer. Untuk mengexplorasi berbagai alogitma CN2 dan mengerti bagaimana pengaturan parameter mempengaruhi proses learning, widget Rule Viewer harus terus open dan dalam penglihatan. Sehingga saat setting CN2 learning algorithm (hasilnya dapat langsung terupdate).

CN2-Viewer-Example1.png

Pemilihan sebuah rule akan mengeluarkan data instance tertentu yang terfilter. Ini dapat dilihat di widget Data Table.

Referensi

Pranala Menarik