0. SECTIONS
-------------
1. Project
2. Dataset
3. Terms of Use
4. Content
5. Method and Processing
6. Changelog

1. PROJECT
------------
Title: Topology Reconfiguration in Distributed Echo State Networks
Dates: Oct 2011 - Sept 2015
Funding organisation: EPSRC & Thales Research UK
Grant no.:

2. DATASET
------------
Title: Topology Reconfiguration in Distributed Echo State Networks Dataset
Description: full parameter-accuracy results for the thesis "Topology Reconfiguration in Distributed Echo State Networks"
Publication Year: 2015

Creator(s) 
Name(s): Daniel Dean
Organisation(s): University of Reading

3. TERMS OF USE
-----------------

This work is licensed by the Creative Commons Attribution license seen at https://creativecommons.org/licenses/by/3.0/ .

4. CONTENT
------------
File listing:
ch3\Hybrid\*            :ESN pool damage compensation using hybrid-type output layer optimisers. File name indicates exact optimiser used
ch3\Least square*       :ESN pool damage compensation using least square-type output layer optimisers. File name indicates exact optimiser used
ch3\Population\*        :ESN pool damage compensation using population based output layer optimisers. File name indicates exact optimiser used

ch4\Hierarchical\NARMA\*:dESN experiments on NARMA30 dataset. file name follows form: hier_{damaged(dynamic)/undamaged(static)}_{training type}_{network config}
ch4\Hierarchical\USPS\* :dESN experiments on USPS dataset. file name follows form: hier_{damaged(dynamic)/undamaged(static)}_{training type}_{network config}
ch4\Recurrent\NARMA\*   :dESN experiments on NARMA30 dataset. file name follows form: recur_{damaged(dynamic)/undamaged(static)}_{training type}_{network config}
ch4\Recurrent\USPS\*    :dESN experiments on USPS dataset. file name follows form: recur_{damaged(dynamic)/undamaged(static)}_{training type}_{network config}

ch5\commfailure.csv     :dESN implementation in a simulation environment. Results for 5x NARAMA30 problems with iterations of discovery.


Additional information
Organisation of files: Files organised by chapter/subsection, as in thesis.
File formats and structure: All *.csv

Required software (including version): Python(x,y) 2.7, OGER library (latest revision as of may 2015), thesis-created tools.


5. METHOD and PROCESSING
--------------------------

See Topology Reconfiguration in Distributed Echo State Networks.


6. CHANGELOG
---------------

22/09/2015: Dataset IP issues resolved. Dataset fully uploaded.
17/09/2015: Created