1. Title: Miba Friction Plate Testing Data
2. Sources:
   (a) Miba frictec, Andreas Promberger, Peter Mitterbauer Str. 1,
       A-4661 Roitham, AUSTRIA, +43 76139020, Andreas.promberger@miba.com
   (b) FH Upper Austria, Gabriel Kronberger, Softwarepark 11,
       A-4232 Hagenberg, AUSTRIA, +43 50804 22320,
       gabriel.kronberger@heuristiclab.com
   (c) March 2017


3. Past Usage:
  - G. Kronberger, M. Kommenda, E. Lughofer, S. Saminger-Platz,
    A. Promberger, F. Nickel, S. Winkler, M. Affenzeller - Robust
    Generalized Fuzzy Modeling and Enhanced Symbolic Regression for
    Modeling Tribological Systems, submitted to Applied Soft
    Computing, 2017

  - E. Lughofer, G. Kronberger, M. Kommenda, S. Saminger-Platz,
    A. Promberger, F. Nickel, S. M. Winkler, M. Affenzeller - Robust
    Fuzzy Modeling and Symbolic Regression for Establishing Accurate
    and Interpretable Prediction Models in Supervising Tribological
    Systems - Proceedings of the 8th International Joint Conference on
    Computational Intelligence, Porto, Portugal, 2016, pp. 51-63

4. Relevant information:
  A set of datasets for regression modelling of
  friction characteristics of friction plate systems. The data stem
  from tests of friction plates with commercial test benches for wet
  friction plate systems. Friction characteristics such as the
  coefficient of friction, wear, and temperatures are measured at
  different loads. The goal is to predict these values given load
  parameters. Data have been kindly provided by Miba frictec company.

5. Number of instances
  A separate file is provided for each target variable. The values of
  the target variable are given in the last column.

  - Cf1: 815 instances
  - Cf2: 2921 instances
  - Cf3: 657 instances
  - Cf4: 649 instances
  - NVH_Rating: 3943 instances
  - Temp1: 656 instances
  - Temp2: 648 instances
  - Wear1: 904 instances
  - Wear2: 902 instances

6. Number of Attributes: 28
  (2 binary, 22 numeric & continuous, and four nominal) Depending on
  the target variable (or file) some of the attributes might be
  constant.

7. Attribute Information
  The first column 'Partition' contains the partition assignment that
  should be used for validation of algorithms. This attribute should
  not be used for modelling. Rows must not be shuffled as in standard
  cross-validation.

  - Source1:      indicates that the row stems from data source one [binary]
  - Source2:      indicates that the row stems from data source two (
                  redundant given Source1) [binary]
  - x1:           an attribute of the testing procedure [numeric, integer]
  - Material_Cat: represents the type of friction material [nominal] 
  - x2, ... ,x16: material attributes  [numeric, continuous] 
  - Material:     represents the friction material [nominal]
  - Grooving:     represents the grooving (surface structure) on the
                  friction plate [nominal]
  - Oil:          represents the oil type [nominal]
  - x17, ... ,x22 load attributes [numeric, continuous]


If you use these data files please use the following reference:
  G. Kronberger, M. Kommenda, E. Lughofer, S. Saminger-Platz,
  A. Promberger, F. Nickel, S. Winkler, M. Affenzeller - Robust
  Generalized Fuzzy Modeling and Enhanced Symbolic Regression for
  Modeling Tribological Systems, submitted to Applied Soft Computing,
  2017