REx
Rule Extractor →
A Logico-Visual Data Analytics Tool
A Logico-Visual Data Analytics Tool
Terms of Use •
Privacy Policy •
Credits •
Feedback •
Versions •
Validation @ 3.138.134.77 •
BibTeX
This Web-application has been tested with Chrome, Firefox, Opera, and Safari. Always use it with the newest versions of these Web-browsers.
By using the Rule Extractor (REx) you are agreeing to these General Terms and Conditions.
All your data are yours.
We only log a hash of your data that you send to our server for main processing.
Back your data up by
Restore your data by
We use no cookie nor device finger-printing.
We only log a hash of your data that you send to our server for main processing.
Back your data up by
- clicking on the backup button and
- copy-pasting the content of the pop-up dialog into a file or message.
Restore your data by
- clicking on the restore button,
- copy-pasting the stored content into the pop-up dialog, and
- clicking the "OK"-button of the dialog window.
We use no cookie nor device finger-printing.
This application has been built with HTML5, CSS3, JavaScript (including lz-string and MathJax), jQuery, and Java; it is physically hosted by firestorm.ch.
The name REx was suggested by Ivan José Varzinczak.
2015/08/06: beta
2015/12/16: + binary-file analyser (BFA)
2016/04/13: + input data table transposition (pre-processing)
2016/04/13: + input data row subtraction (pre-processing)
2016/04/13: + input data sign retention (pre-processing)
2016/04/19: + hypothesis calculator (post-processing)
2016/04/19: + distance calculator (post-processing)
2016/11/29: graphical optimisation (zeichenfabrik.ch)
2017/03/11: BFA + {summaries, details zip}
2017/07/30: + complementarity calculator
@misc{TheRExWebApp,
Author = {S. Kramer},
Howpublished = {\url{https://www.rule-extractor.ch}},
Title = {{Rule Extractor} {(REx)} --- {A} {L}ogico-{V}isual {D}ata-{A}nalytics {T}ool},
Year = {2015}}
REx offers
- an optimal algorithm for the automatic extraction as a rulebase and structured visualisation as a rule diagram of a sound, thus consistent, and complete set of constructive predicate reasoning rules (implicational invariants) from arbitrary tabular data;
- a lightweight algorithm for a possible pre-processing of the data; and
-
a suit of three post-processing algorithms for
- the calculation of the complete set of consequences (deductive closure) that can be deduced from the extracted rules under a user-input finite set of elementary hypotheses by means of modus ponens,
- the set-based comparison of and operation on rulebases, and
- the model-checking of an existing input data table and rulebase including the visualisation of the check as a satisfaction table.
As a desirable side effect, Algorithm 1 also extracts and visualises a complete set of quantitatively qualified, fuzzy (not invariant) reasoning rules from its input data.