Note
This webpage summarizes information about the demonstration of the Siren for mining redescriptions with trees, submitted to the 2015 IEEE International Conference on Data Mining, Atlantic City, NJ, USA on November 14-17, 2015.
Esther Galbrun and Pauli Miettinen. Mining predictive Redescriptions with Trees. Submitted to ICDM. 2015. Original paper.
More details can be found on the main Siren webpage or in the user guide.
Redescription mining is a powerful data analysis tool that aims at finding alternative descriptions of the same entities.
For example, in biology, an important task is to identify the bioclimatic constraints that allow some species to survive, that is, to describe geographical regions in terms of both their bioclimatic conditions and the fauna that inhabit them.
Siren is a tool for interactive mining and visualization of redescriptions. We integrated tree-based redescription mining algorithms, allowing to find redescriptions that generalize well.
Siren is a multi-platform software implemented in Python.
Siren and ReReMi are licensed under the Apache License, Version 2.0.
In Siren, data include:
Obviously, this is required.
Data can be imported to Siren via the interface menu .
Data can be imported into Siren as CSV files. The program expects a pair of files, one for either side in character-separated values, as can be imported and exported to and from spreadsheet programms, for instance.
In particular, the data can stored as a table with one column for each variable and one row each entity. The first row can contain the names of the variables. The entities names can be included as columns named ids. Similarly the coordinates can be included as a pair of columns named longitudes and latitudes, respectively.
There are various strategies for mining redescriptions mining. We integrated tree-based algorithms to the Siren interface to allow mining redescriptions that generalize better, than, for instance redescriptions mine with the greedy ReReMi algorithm. For more details, check the references section.
The first algorithm introduced for redescription mining was actually based on alternating between constructing CARTs and hence was called the CARTWheels algorithm.
See the little slideshow below to understand how redescriptions are constructed with this approach and read the corresponding publication in the references section for more details.
An alternative method for constructing CARTs is to build them layer by layer, we call this method the layered trees.
Finally the third method available in Siren construct queries by refining the CART branches separately, we call this method the split trees.
We provide a prepared dataset about the Finnish 2011 parliamentary elections. Get the data (non-geospatial), try out Siren and learn about the finnish political scene! (More details on the main webpage.)
To illustrate the use of Siren, we present example use-cases from different application domains.
One use-case concerns niche-finding, i.e. the problem of finding species’ bioclimatic envelope, an important task in biology.
| [A] | Esther Galbrun and Pauli Miettinen. Siren: an interactive tool for mining and visualizing geospatial redescriptions. In KDD, 1544–1547. ACM, 2012. Preprint, Poster. |
| [B] | Esther Galbrun and Pauli Miettinen. From black and white to full color: extending redescription mining outside the Boolean world. Statistical Analysis and Data Mining, 5(4):284–303, 2012. Preprint. |
| [C] | Esther Galbrun. Methods for Redescription Mining. PhD thesis, University of Helsinki, 2014. http://urn.fi/URN:ISBN:978-952-10-9431-6 . |
| [D] | Tetiana Zinchenko. Redescription Mining Over non-Binary Data Sets Using Decision Trees. MSc thesis, University of Saarland and Max-Planck Institute for Informatics, 2014. http://www.mpi-inf.mpg.de/~pmiettin/papers/zinchenko15redescription.pdf . |
| [E] | Naren Ramakrishnan, Deept Kumar, Bud Mishra, Malcolm Potts, and Richard F Helm. Turning CARTwheels: An Alternating Algorithm for Mining Redescriptions. In KDD, 266–275. ACM, 2004. |
