From 63abdc782c1911c89cab60a7f472552826185b3e Mon Sep 17 00:00:00 2001
From: emeric <itmfr@yahoo.fr>
Date: Sun, 9 Dec 2018 15:20:29 +0100
Subject: [PATCH] Adding a weight for each input dimension

---
 src/classifier/SOM.cpp             | 23 +++++++++++++++++------
 src/classifier/SOM.hpp             |  8 ++++++--
 tools/classifier/LmsClassifier.cpp |  1 +
 3 files changed, 24 insertions(+), 8 deletions(-)

diff --git a/src/classifier/SOM.cpp b/src/classifier/SOM.cpp
index d495b49e..32e1c67b 100644
--- a/src/classifier/SOM.cpp
+++ b/src/classifier/SOM.cpp
@@ -50,13 +50,16 @@ defaultLearningFactor(Network::Progress progress)
 }
 
 InputVector::value_type
-euclidianSquareDistance(const InputVector& a, const InputVector& b)
+euclidianSquareDistance(const InputVector& a, const InputVector& b, const InputVector& weights)
 {
+	checkSameDimensions(a, b);
+	checkSameDimensions(a, weights);
+
 	InputVector::value_type res = 0;
 
 	for (std::size_t i = 0; i < a.size(); ++i)
 	{
-		res += (a[i] - b[i]) * (a[i] - b[i]);
+		res += (a[i] - b[i]) * (a[i] - b[i]) * weights[i];
 	}
 
 	return res;
@@ -159,6 +162,7 @@ Network::Network(std::size_t width, std::size_t height, std::size_t inputDimCoun
 : _width(width),
 _height(height),
 _inputDimCount(inputDimCount),
+_weights(inputDimCount, static_cast<InputVector::value_type>(1)),
 _distanceFunc(euclidianSquareDistance),
 _learningFactorFunc(defaultLearningFactor),
 _neighborhoodFunc(defaultNeighborhoodFunc)
@@ -180,6 +184,13 @@ _neighborhoodFunc(defaultNeighborhoodFunc)
 	}
 }
 
+void
+Network::setDataWeights(const InputVector& weights)
+{
+	checkSameDimensions(weights, _inputDimCount);
+
+	_weights = weights;
+}
 
 InputVector&
 Network::getRefVector(std::size_t x, std::size_t y)
@@ -217,7 +228,7 @@ Network::getClosestRefVector(const InputVector& data) const
 	auto it = std::min_element(_refVectors.begin(), _refVectors.end(),
 			[&](const auto& a, const auto& b)
 			{
-				return (_distanceFunc(a, data) < _distanceFunc(b, data));
+				return (_distanceFunc(a, data, _weights) < _distanceFunc(b, data, _weights));
 			});
 
 	auto index = std::distance(_refVectors.begin(), it);
@@ -271,12 +282,12 @@ Network::train(const std::vector<InputVector>& inputData, std::size_t nbIteratio
 		inputDataShuffled.push_back(&input);
 	}
 
-	std::random_device randomDevice;
-	std::mt19937 generator(randomDevice());
+	auto now = std::chrono::system_clock::now();
+	std::mt19937 randGenerator(std::chrono::duration_cast<std::chrono::milliseconds>(now.time_since_epoch()).count());
 
 	for (std::size_t i = 0; i < nbIterations; ++i)
 	{
-		std::shuffle(inputDataShuffled.begin(), inputDataShuffled.end(), generator);
+		std::shuffle(inputDataShuffled.begin(), inputDataShuffled.end(), randGenerator);
 
 		for (auto input : inputDataShuffled)
 		{
diff --git a/src/classifier/SOM.hpp b/src/classifier/SOM.hpp
index f9346f93..2d60ff47 100644
--- a/src/classifier/SOM.hpp
+++ b/src/classifier/SOM.hpp
@@ -52,6 +52,9 @@ class Network
 
 		Network(std::size_t width, std::size_t height, std::size_t inputDimCount);
 
+		// Set weight for each dimension (default is 1 for each weight)
+		void setDataWeights(const InputVector& weights);
+
 		// data must be normalized
 		void train(const std::vector<InputVector>& dataSamples, std::size_t nbIterations);
 
@@ -64,7 +67,7 @@ class Network
 		// i is the current iteration
 		// refVector(i+1) = refVector(i) + LearningFactor(i) * NeighborhoodFunc(i) * (MatchingRefVector - refVector)
 
-		using DistanceFunc = std::function<InputVector::value_type(const InputVector&, const InputVector&)>;
+		using DistanceFunc = std::function<InputVector::value_type(const InputVector& /* a */, const InputVector& /* b */, const InputVector& /* weights */)>;
 		void setDistanceFunc(DistanceFunc distanceFunc);
 
 		struct Progress
@@ -91,7 +94,8 @@ class Network
 		std::size_t _height;
 		std::size_t _inputDimCount;
 
-		std::vector<InputVector> _refVectors; // indexed reference vectors
+		InputVector _weights;
+		std::vector<InputVector> _refVectors; // reference vectors
 
 		DistanceFunc _distanceFunc;
 		LearningFactorFunc _learningFactorFunc;
diff --git a/tools/classifier/LmsClassifier.cpp b/tools/classifier/LmsClassifier.cpp
index 36568e42..620cd956 100644
--- a/tools/classifier/LmsClassifier.cpp
+++ b/tools/classifier/LmsClassifier.cpp
@@ -29,6 +29,7 @@ int main(int argc, char *argv[])
 		{ 80, -1 },
 		{ 80, -0.75 },
 		{ 240, 0.5 },
+		{ 240, -0.5 },
 		{ 120, -0.5 },
 		{ 140, -0.5 },
 	};