rmontanana/mdlp
1
1
Fork 0
mirror of https://github.com/rmontanana/mdlp.git synced 2025-08-16 07:55:58 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add max_cutpoints Hyperparameter

Browse Source
This commit is contained in:
Ricardo Montañana Gómez
2023-03-13 01:17:04 +01:00
parent ed784736ca
commit ffb8df4d1c
7 changed files with 105 additions and 41 deletions
Download Patch File Download Diff File Expand all files Collapse all files

43
CPPFImdlp.cpp
View File

@@ -3,25 +3,42 @@
#include <set>
#include <cmath>
#include <limits>
#include <cmath>
#include "CPPFImdlp.h"
#include "Metrics.h"
namespace mdlp {
CPPFImdlp::CPPFImdlp():depth(0), max_depth(numeric_limits<int>::max()), min_length(3), indices(indices_t()), X(samples_t()), y(labels_t()),
metrics(Metrics(y, indices))
CPPFImdlp::CPPFImdlp():min_length(3), depth(0), max_depth(numeric_limits<int>::max()), proposed_cuts(0),
indices(indices_t()), X(samples_t()), y(labels_t()),
metrics(Metrics(y, indices)), num_cut_points(numeric_limits<size_t>::max())
{
}
CPPFImdlp::CPPFImdlp(size_t min_length_, int max_depth_): depth(0), max_depth(max_depth_), min_length(min_length_), indices(indices_t()), X(samples_t()), y(labels_t()),
metrics(Metrics(y, indices))
CPPFImdlp::CPPFImdlp(size_t min_length_, int max_depth_, float proposed): min_length(min_length_), depth(0),
max_depth(max_depth_), proposed_cuts(proposed), indices(indices_t()), X(samples_t()), y(labels_t()),
metrics(Metrics(y, indices)), num_cut_points(numeric_limits<size_t>::max())
{
}
CPPFImdlp::~CPPFImdlp() = default;
CPPFImdlp& CPPFImdlp::fit(samples_t& X_, labels_t& y_)
size_t CPPFImdlp::compute_max_num_cut_points()
{
// Set the actual maximum number of cut points as a number or as a percentage of the number of samples
if (proposed_cuts == 0) {
return numeric_limits<size_t>::max();
}
if (proposed_cuts < 0 || proposed_cuts > X.size()) {
throw invalid_argument("wrong proposed num_cuts value");
}
if (proposed_cuts < 1)
return (int)round(X.size() * proposed_cuts);
return (int)proposed_cuts;
}
void CPPFImdlp::fit(samples_t& X_, labels_t& y_)
{
X = X_;
y = y_;
num_cut_points = compute_max_num_cut_points();
depth = 0;
cutPoints.clear();
if (X.size() != y.size()) {
@@ -39,7 +56,6 @@ namespace mdlp {
indices = sortIndices(X_, y_);
metrics.setData(y, indices);
computeCutPoints(0, X.size(), 1);
return *this;
}
pair<precision_t, size_t> CPPFImdlp::valueCutPoint(size_t start, size_t cut, size_t end)
@@ -75,6 +91,8 @@ namespace mdlp {
{
size_t cut;
pair<precision_t, size_t> result;
if (cutPoints.size() == num_cut_points)
return;
// Check if the interval length and the depth are Ok
if (end - start < min_length || depth_ > max_depth)
return;
@@ -158,15 +176,8 @@ namespace mdlp {
cutPoints_t CPPFImdlp::getCutPoints()
{
// Remove duplicates and sort
cutPoints_t output(cutPoints.size());
set<precision_t> s;
unsigned size = cutPoints.size();
for (unsigned i = 0; i < size; i++)
s.insert(cutPoints[i]);
output.assign(s.begin(), s.end());
sort(output.begin(), output.end());
return output;
sort(cutPoints.begin(), cutPoints.end());
return cutPoints;
}
int CPPFImdlp::get_depth()
{