/**
 * @file sample-ml-app.cpp
 * @author German Lancioni
 *
 * Sample app based on the Windows ML App Tutorial.
 * Quickly shows how to create a machine learning app using mlpack/C++.
 *
 * mlpack is free software; you may redistribute it and/or modify it under the
 * terms of the 3-clause BSD license.  You should have received a copy of the
 * 3-clause BSD license along with mlpack.  If not, see
 * http://www.opensource.org/licenses/BSD-3-Clause for more information.
 */
#include "stdafx.h"

using namespace arma;
using namespace mlpack;
using namespace std;

/*
This sample app covers a brief end-to-end ML workflow
as you would do in a real-life application. Including:
- Loading and preparing a dataset
- Training (Random Forest as example)
- Computing the training accuracy
- Cross-Validation using K-Fold
- Metrics gathering (accuracy, precision, recall, F1)
- Saving the trained model
- Loading the model
- Classifying a new sample

Assumptions:
- No labels normalization required
*/
int main()
{
	cout << "[SAMPLE:BEGIN]";

	// (1) Load the dataset
	cout << "\nLoading dataset...";

	mat dataset;

	// CSV is loaded transposed (columns are samples, rows are dimensions)
	bool loaded = mlpack::data::Load("data/german.csv", dataset);
	if (!loaded)
		return -1;

	Row<size_t> labels;

	// Extract the labels from the last dimension of the training set
	labels = conv_to<Row<size_t>>::from(dataset.row(dataset.n_rows - 1));

	// Remove the labels from the training set
	dataset.shed_row(dataset.n_rows - 1);

	// (2) Training
	cout << "\nTraining...";
	const size_t numClasses = 2;
	const size_t minimumLeafSize = 5;
	const size_t numTrees = 10;

	RandomForest<GiniGain, RandomDimensionSelect> rf;

	rf = RandomForest<GiniGain, RandomDimensionSelect>(dataset, labels,
		numClasses, numTrees, minimumLeafSize);

	Row<size_t> predictions;
	rf.Classify(dataset, predictions);

	const size_t correct = arma::accu(predictions == labels);

	cout << "\nTraining Accuracy: " << (double(correct) / double(labels.n_elem));

	// (3) Cross-Validation
	cout << "\nCross-Validating...";
	const size_t k = 10;

	KFoldCV<RandomForest<GiniGain, RandomDimensionSelect>, Accuracy> cv(k,
		dataset, labels, numClasses);

	double cvAcc = cv.Evaluate(numTrees, minimumLeafSize);
	cout << "\nKFoldCV Accuracy: " << cvAcc;

	double cvPrecision = Precision<Binary>::Evaluate(rf, dataset, labels);
	cout << "\nPrecision: " << cvPrecision;

	double cvRecall = Recall<Binary>::Evaluate(rf, dataset, labels);
	cout << "\nRecall: " << cvRecall;

	double cvF1 = F1<Binary>::Evaluate(rf, dataset, labels);
	cout << "\nF1: " << cvF1;

	// (4) Save the model
	cout << "\nSaving model...";
	mlpack::data::Save("mymodel.xml", "model", rf, false);

	// (5) Load the model
	cout << "\nLoading model...";
	mlpack::data::Load("mymodel.xml", "model", rf);

	// (6) Classify a new sample
	cout << "\nClassifying a new sample...";
	// Should classify as "1"
	mat sample("2 12 2 13 1 2 2 1 3 24 3 1 1 1 1 1 0 1 0 1 0 0 0");

	mat probabilities;

	rf.Classify(sample, predictions, probabilities);

	u64 result = predictions.at(0);

	cout << "\nClassification result: " << result << " , Probabilities: " <<
		probabilities.at(0) << "/" << probabilities.at(1);

	cout << "\n[SAMPLE:END]\n";
	return 0;
}