From 7f97812915c194a3c42857e576dc0bf2a2fde06b Mon Sep 17 00:00:00 2001
From: Nuno Fachada <faken@fakenmc.com>
Date: Sat, 30 May 2020 16:59:09 +0100
Subject: [PATCH] Update citation reference

---
 README.md      | 23 +++++++++++++++++++----
 generateData.m | 24 ++++++++++++++----------
 2 files changed, 33 insertions(+), 14 deletions(-)

diff --git a/README.md b/README.md
index 7b1a5e5..e3f48be 100644
--- a/README.md
+++ b/README.md
@@ -1,5 +1,6 @@
 [![Latest release](https://img.shields.io/github/release/fakenmc/generateData.svg)](https://github.com/fakenmc/generateData/releases)
 [![MIT Licence](https://img.shields.io/badge/license-MIT-yellowgreen.svg)](https://opensource.org/licenses/MIT/)
+[![View Generate Data for Clustering on File Exchange](https://www.mathworks.com/matlabcentral/images/matlab-file-exchange.svg)](https://www.mathworks.com/matlabcentral/fileexchange/37435-generate-data-for-clustering)
 
 # generateData
 
@@ -100,14 +101,28 @@ rand("state", 123);
 randn("state", 123);
 ```
 
+## Previous behaviors and reproducibility of results
+
+Before [v2.0.0](https://github.com/fakenmc/generateData/tree/v2.0.0), lines
+supporting clusters were parameterized with slopes instead of angles. We found
+this caused difficulties when choosing line orientation, thus the change to
+angles, which are much easier to work with.
+Version [v1.3.0](https://github.com/fakenmc/generateData/tree/v1.3.0) still
+uses slopes, for those who prefer this behavior.
+
+For reproducing results in studies published before May 2020, use version
+[v1.2.0](https://github.com/fakenmc/generateData/tree/v1.2.0) instead.
+Subsequent versions were optimized in a way that changed the order in which
+the required random values are generated, thus producing slightly different
+results.
+
 ## Reference
 
 If you use this function in your work, please cite the following reference:
 
-- Fachada, N., Figueiredo, M.A.T., Lopes, V.V., Martins, R.C., Rosa,
-A.C., [Spectrometric differentiation of yeast strains using minimum volume
-increase and minimum direction change clustering criteria](http://www.sciencedirect.com/science/article/pii/S0167865514000889),
-Pattern Recognition Letters, vol. 45, pp. 55-61 (2014), doi: http://dx.doi.org/10.1016/j.patrec.2014.03.008
+- Fachada, N., & Rosa, A. C. (2020).
+[generateData—A 2D data generator](https://doi.org/10.1016/j.simpa.2020.100017).
+Software Impacts, 4:100017. doi: [10.1016/j.simpa.2020.100017](https://doi.org/10.1016/j.simpa.2020.100017)
 
 ## License
 
diff --git a/generateData.m b/generateData.m
index 94b9477..e17fd0a 100644
--- a/generateData.m
+++ b/generateData.m
@@ -11,11 +11,11 @@
         totalPoints, ...
         varargin ...
     )
-% GENERATEDATA Generates 2D data for clustering. Data is created along 
+% GENERATEDATA Generates 2D data for clustering. Data is created along
 %              straight lines, which can be more or less parallel
 %              depending on the angleStd parameter.
 %
-% [data clustPoints idx centers angles lengths] = 
+% [data clustPoints idx centers angles lengths] =
 %    GENERATEDATA(angleMean, angleStd, numClusts, xClustAvgSep, ...
 %                 yClustAvgSep, lengthMean, lengthStd, lateralStd, ...
 %                 totalPoints, ...)
@@ -31,7 +31,7 @@
 %                Line lengths are drawn from the folded normal
 %                distribution.
 %    lengthStd - Standard deviation of line lengths.
-%   lateralStd - Cluster "fatness", i.e., the standard deviation of the 
+%   lateralStd - Cluster "fatness", i.e., the standard deviation of the
 %                distance from each point to its projection on the
 %                line. The way this distance is obtained is controlled by
 %                the optional 'pointOffset' parameter.
@@ -64,9 +64,9 @@
 %                of each point.
 %      centers - Matrix (numClusts x 2) containing centers from where
 %                clusters were generated.
-%       angles - Vector (numClusts x 1) containing the effective angles 
+%       angles - Vector (numClusts x 1) containing the effective angles
 %                of the lines used to generate clusters.
-%      lengths - Vector (numClusts x 1) containing the effective lengths 
+%      lengths - Vector (numClusts x 1) containing the effective lengths
 %                of the lines used to generate clusters.
 %
 % ----------------------------------------------------------
@@ -74,8 +74,8 @@
 %
 %   [data cp idx] = GENERATEDATA(pi / 2, pi / 8, 5, 15, 15, 5, 1, 2, 200);
 %
-% This creates 5 clusters with a total of 200 points, with a mean angle 
-% of pi/2 (std=pi/8), separated in average by 15 units in both x and y 
+% This creates 5 clusters with a total of 200 points, with a mean angle
+% of pi/2 (std=pi/8), separated in average by 15 units in both x and y
 % directions, with mean length of 5 units (std=1) and a "fatness" or
 % spread of 2 units.
 %
@@ -84,8 +84,12 @@
 %   scatter(data(:, 1), data(:, 2), 8, idx);
 
 % Copyright (c) 2012-2020 Nuno Fachada
-% Distributed under the MIT License (See accompanying file LICENSE or copy 
+% Distributed under the MIT License (See accompanying file LICENSE or copy
 % at http://opensource.org/licenses/MIT)
+%
+% Reference:
+% Fachada, N., & Rosa, A. C. (2020). generateData—A 2D data generator.
+% Software Impacts, 4:100017. doi: 10.1016/j.simpa.2020.100017
 
 % Known distributions for sampling points along lines
 pointDists = {'unif', 'norm'};
@@ -225,7 +229,7 @@
         % each point
         perpAngles = angles(i) + sign(points_dist) * pi / 2;
         perpVecs = [cos(perpAngles) sin(perpAngles)];
-        
+
         % Set vector magnitudes
         perpVecs = abs(points_dist) .* perpVecs;
 
@@ -253,4 +257,4 @@
 
     % Update idx
     idx(cumSumPoints(i) + 1 : cumSumPoints(i + 1)) = i;
-end;
\ No newline at end of file
+end;