File size: 13,731 Bytes
41cacc0 9e66e53 41cacc0 9e66e53 41cacc0 9e66e53 41cacc0 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 |
// --- GLOBAL VARIABLES ---
let dataChart, rocChart, metricsChart;
const N_SAMPLES_PER_CLASS = 100;
const metricExplanations = {
'AUC': {
description: "Measures the model's ability to distinguish between positive and negative classes. It represents the probability that a random positive instance is ranked higher than a random negative instance.",
range: "Ranges from 0 (worst) to 1 (best). 0.5 is random chance.",
formula: "Area Under the ROC Curve"
},
'Accuracy': {
description: "The proportion of all predictions that are correct. It's a general measure of the model's performance.",
range: "Ranges from 0 (worst) to 1 (best).",
formula: "(TP + TN) / (TP + TN + FP + FN)"
},
'Precision': {
description: "Of all the positive predictions made by the model, how many were actually positive. High precision indicates a low false positive rate.",
range: "Ranges from 0 (worst) to 1 (best).",
formula: "TP / (TP + FP)"
},
'Recall': {
description: "Of all the actual positive instances, how many did the model correctly identify. Also known as Sensitivity or True Positive Rate.",
range: "Ranges from 0 (worst) to 1 (best).",
formula: "TP / (TP + FN)"
},
'Specificity': {
description: "Of all the actual negative instances, how many did the model correctly identify. Also known as True Negative Rate.",
range: "Ranges from 0 (worst) to 1 (best).",
formula: "TN / (TN + FP)"
},
'F1-Score': {
description: "The harmonic mean of Precision and Recall. It provides a single score that balances both concerns, useful for imbalanced classes.",
range: "Ranges from 0 (worst) to 1 (best).",
formula: "2 * (Precision * Recall) / (Precision + Recall)"
}
};
// --- DATA GENERATION ---
function randomGaussian(mean = 0, stdDev = 1) {
let u = 0, v = 0;
while (u === 0) u = Math.random();
while (v === 0) v = Math.random();
return mean + stdDev * Math.sqrt(-2.0 * Math.log(u)) * Math.cos(2.0 * Math.PI * v);
}
function generateData(separation, stdDev) {
const data = [], labels = [];
for (let i = 0; i < N_SAMPLES_PER_CLASS; i++) { data.push({ x: randomGaussian(-separation / 2, stdDev), y: randomGaussian(0, stdDev) }); labels.push(0); }
for (let i = 0; i < N_SAMPLES_PER_CLASS; i++) { data.push({ x: randomGaussian(separation / 2, stdDev), y: randomGaussian(0, stdDev) }); labels.push(1); }
return { data, labels };
}
// --- CLASSIFIER: GAUSSIAN NAIVE BAYES ---
class GaussianNB {
fit(X, y) {
const classes = [...new Set(y)];
this.classes = classes;
this.params = {};
for (const cls of classes) {
const X_cls = X.filter((_, i) => y[i] === cls);
const mean_x = X_cls.reduce((a, b) => a + b.x, 0) / X_cls.length;
const mean_y = X_cls.reduce((a, b) => a + b.y, 0) / X_cls.length;
this.params[cls] = {
prior: X_cls.length / X.length,
mean: [mean_x, mean_y],
variance: [Math.max(1e-9, X_cls.reduce((a, b) => a + Math.pow(b.x - mean_x, 2), 0) / X_cls.length), Math.max(1e-9, X_cls.reduce((a, b) => a + Math.pow(b.y - mean_y, 2), 0) / X_cls.length)]
};
}
}
_pdf(x, mean, variance) { const exponent = Math.exp(-Math.pow(x - mean, 2) / (2 * variance)); return (1 / Math.sqrt(2 * Math.PI * variance)) * exponent; }
predict_proba(X) {
return X.map(point => {
const posteriors = {};
for (const cls of this.classes) {
const prior = Math.log(this.params[cls].prior);
const likelihood_x = Math.log(this._pdf(point.x, this.params[cls].mean[0], this.params[cls].variance[0]));
const likelihood_y = Math.log(this._pdf(point.y, this.params[cls].mean[1], this.params[cls].variance[1]));
posteriors[cls] = prior + likelihood_x + likelihood_y;
}
const max_posterior = Math.max(...Object.values(posteriors));
const exps = Object.fromEntries(Object.entries(posteriors).map(([k, v]) => [k, Math.exp(v - max_posterior)]));
const sum_exps = Object.values(exps).reduce((a, b) => a + b);
return exps[1] / sum_exps;
});
}
}
// --- METRICS CALCULATIONS ---
function calculateRocAndAuc(labels, scores) {
const pairs = labels.map((label, i) => ({ label, score: scores[i] }));
pairs.sort((a, b) => b.score - a.score);
let tp = 0, fp = 0;
const total_pos = labels.filter(l => l === 1).length;
const total_neg = labels.length - total_pos;
if (total_pos === 0 || total_neg === 0) return { rocPoints: [{ x: 0, y: 0 }, { x: 1, y: 1 }], auc: 0.5 };
const rocPoints = [{ x: 0, y: 0 }];
let auc = 0, prev_tpr = 0, prev_fpr = 0;
for (const pair of pairs) {
if (pair.label === 1) tp++; else fp++;
const tpr = tp / total_pos;
const fpr = fp / total_neg;
auc += (tpr + prev_tpr) / 2 * (fpr - prev_fpr);
rocPoints.push({ x: fpr, y: tpr });
prev_tpr = tpr; prev_fpr = fpr;
}
return { rocPoints, auc };
}
function getConfusionMatrix(labels, scores, threshold) {
let vp = 0, fp = 0, vn = 0, fn = 0;
labels.forEach((label, i) => {
const prediction = scores[i] >= threshold ? 1 : 0;
if (prediction === 1 && label === 1) vp++;
else if (prediction === 1 && label === 0) fp++;
else if (prediction === 0 && label === 0) vn++;
else if (prediction === 0 && label === 1) fn++;
});
return { vp, fp, vn, fn };
}
function drawConfusionMatrix(canvasId, vp, fp, vn, fn) {
const canvas = document.getElementById(canvasId);
const ctx = canvas.getContext('2d');
const w = canvas.width, h = canvas.height;
ctx.clearRect(0, 0, w, h);
const margin = 50, gridW = w - margin, gridH = h - margin, cellW = gridW / 2, cellH = gridH / 2;
const max_val = Math.max(vp, fp, vn, fn);
const baseColor = [8, 48, 107];
const cells = [{ label: 'TN', value: vn, x: 0, y: cellH }, { label: 'FP', value: fp, x: cellW, y: cellH }, { label: 'FN', value: fn, x: 0, y: 0 }, { label: 'TP', value: vp, x: cellW, y: 0 }];
cells.forEach(cell => {
const intensity = max_val > 0 ? cell.value / max_val : 0;
ctx.fillStyle = `rgba(${baseColor[0]}, ${baseColor[1]}, ${baseColor[2]}, ${intensity})`;
ctx.fillRect(margin + cell.x, cell.y, cellW, cellH);
ctx.fillStyle = intensity > 0.5 ? 'white' : 'black';
ctx.textAlign = 'center'; ctx.textBaseline = 'middle';
ctx.font = 'bold 20px Segoe UI'; ctx.fillText(cell.label, margin + cell.x + cellW / 2, cell.y + cellH / 2 - 12);
ctx.font = '18px Segoe UI'; ctx.fillText(cell.value, margin + cell.x + cellW / 2, cell.y + cellH / 2 + 12);
});
ctx.fillStyle = '#333'; ctx.font = 'bold 14px Segoe UI';
ctx.fillText('Negative', margin + cellW / 2, gridH + 20);
ctx.fillText('Positive', margin + cellW + cellW / 2, gridH + 20);
ctx.save();
ctx.translate(20, gridH / 2);
ctx.rotate(-Math.PI / 2);
ctx.textAlign = 'center';
ctx.textBaseline = 'middle';
ctx.fillText('Positive', -cellH / 2, 0);
ctx.fillText('Negative', cellH / 2, 0);
ctx.restore();
}
// --- UI UPDATE ---
function updateApplication() {
const separation = parseFloat(document.getElementById('separationSlider').value);
const stdDev = parseFloat(document.getElementById('stdDevSlider').value);
document.getElementById('separationValue').textContent = separation.toFixed(1);
document.getElementById('stdDevValue').textContent = stdDev.toFixed(1);
const { data, labels } = generateData(separation, stdDev);
const model = new GaussianNB();
model.fit(data, labels);
const scores = model.predict_proba(data);
const { rocPoints, auc } = calculateRocAndAuc(labels, scores);
const { vp, fp, vn, fn } = getConfusionMatrix(labels, scores, 0.5);
const total = vp + fp + vn + fn;
const precision = (vp + fp) > 0 ? vp / (vp + fp) : 0;
const recall = (vp + fn) > 0 ? vp / (vp + fn) : 0;
const specificity = (vn + fp) > 0 ? vn / (vn + fp) : 0;
const f1score = (precision + recall) > 0 ? 2 * (precision * recall) / (precision + recall) : 0;
const accuracy = total > 0 ? (vp + vn) / total : 0;
drawConfusionMatrix('matrixChart', vp, fp, vn, fn);
dataChart.data.datasets[0].data = data.filter((_, i) => labels[i] === 0);
dataChart.data.datasets[1].data = data.filter((_, i) => labels[i] === 1);
dataChart.update('none');
rocChart.data.datasets[0].data = rocPoints;
rocChart.update('none');
metricsChart.data.datasets[0].data = [auc, accuracy, precision, recall, specificity, f1score];
metricsChart.update('none');
}
// --- INITIALIZATION ---
const customDatalabelsPlugin = {
id: 'customDatalabels',
afterDatasetsDraw: (chart) => {
const ctx = chart.ctx;
ctx.save();
ctx.font = 'bold 12px Segoe UI';
ctx.fillStyle = 'white';
ctx.textAlign = 'center';
chart.data.datasets.forEach((dataset, i) => {
const meta = chart.getDatasetMeta(i);
meta.data.forEach((bar, index) => {
const data = dataset.data[index];
if (bar.height > 15) {
ctx.textBaseline = 'bottom';
ctx.fillText(data.toFixed(3), bar.x, bar.y + bar.height - 5);
}
});
});
ctx.restore();
}
};
function initCharts() {
const dataCtx = document.getElementById('dataChart').getContext('2d');
dataChart = new Chart(dataCtx, { type: 'scatter', data: { datasets: [{ label: 'Negative Class', data: [], backgroundColor: '#0D47A1' }, { label: 'Positive Class', data: [], backgroundColor: '#B71C1C' }] }, options: { responsive: true, maintainAspectRatio: false, animation: { duration: 0 } } });
const rocCtx = document.getElementById('rocChart').getContext('2d');
rocChart = new Chart(rocCtx, { type: 'scatter', data: { datasets: [{ label: 'ROC Curve', data: [], borderColor: '#0D47A1', backgroundColor: 'transparent', showLine: true, pointRadius: 0, borderWidth: 3 }, { label: 'Chance Line', data: [{ x: 0, y: 0 }, { x: 1, y: 1 }], borderColor: '#666', showLine: true, pointRadius: 0, borderDash: [5, 5] }] }, options: { responsive: true, maintainAspectRatio: false, animation: { duration: 0 }, scales: { x: { min: 0, max: 1, title: { display: true, text: 'False Positive Rate' } }, y: { min: 0, max: 1, title: { display: true, text: 'True Positive Rate' } } } } });
const metricsCtx = document.getElementById('metricsChart').getContext('2d');
metricsChart = new Chart(metricsCtx, {
type: 'bar',
data: { labels: ['AUC', 'Accuracy', 'Precision', 'Recall', 'Specificity', 'F1-Score'], datasets: [{ data: [], backgroundColor: ['#673AB7', '#009688', '#1E88E5', '#388E3C', '#FB8C00', '#9C27B0'] }] },
plugins: [customDatalabelsPlugin],
options: {
responsive: true,
maintainAspectRatio: false,
indexAxis: 'x',
animation: { duration: 0 },
plugins: {
legend: { display: false },
tooltip: {
enabled: true,
backgroundColor: 'rgba(255, 255, 255, 0.95)',
titleColor: '#000',
bodyColor: '#000',
borderColor: '#555',
borderWidth: 1,
padding: 15,
displayColors: false,
callbacks: {
label: function (context) {
const label = context.chart.data.labels[context.dataIndex];
const value = context.raw.toFixed(3);
const explanation = metricExplanations[label];
let tooltipText = [`${label}: ${value}`];
if (explanation) {
tooltipText.push('');
const roleLines = `Role: ${explanation.description}`.match(/.{1,50}(\s|$)/g) || [];
roleLines.forEach(line => tooltipText.push(line.trim()));
tooltipText.push(`Range: ${explanation.range}`);
tooltipText.push(`Formula: ${explanation.formula}`);
}
return tooltipText;
}
}
}
},
scales: { y: { beginAtZero: true, max: 1 } }
}
});
}
function makeDraggable(element, handle) {
let pos1 = 0, pos2 = 0, pos3 = 0, pos4 = 0;
handle.onmousedown = (e) => { e.preventDefault(); pos3 = e.clientX; pos4 = e.clientY; document.onmouseup = closeDragElement; document.onmousemove = elementDrag; };
const elementDrag = (e) => { e.preventDefault(); pos1 = pos3 - e.clientX; pos2 = pos4 - e.clientY; pos3 = e.clientX; pos4 = e.clientY; element.style.top = (element.offsetTop - pos2) + "px"; element.style.left = (element.offsetLeft - pos1) + "px"; };
const closeDragElement = () => { document.onmouseup = null; document.onmousemove = null; };
}
window.addEventListener('load', function () {
initCharts();
const sliders = ['separationSlider', 'stdDevSlider'];
sliders.forEach(id => { document.getElementById(id).addEventListener('input', updateApplication); });
if (window.innerWidth > 1200) { makeDraggable(document.getElementById('floatingControls'), document.getElementById('controlsTitle')); }
updateApplication();
});
|