4f5578c
| 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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 | <!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<meta name="viewport" content="width=device-width, initial-scale=1.0">
<title>Feature Encoding Methods Visualizer</title>
<style>
* {
margin: 0;
padding: 0;
box-sizing: border-box;
}
body {
font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
padding: 20px;
min-height: 100vh;
}
.container {
max-width: 1200px;
margin: 0 auto;
background: white;
border-radius: 15px;
padding: 30px;
box-shadow: 0 20px 60px rgba(0,0,0,0.3);
}
h1 {
text-align: center;
color: #667eea;
margin-bottom: 30px;
font-size: 2.5em;
}
.method-selector {
display: flex;
flex-wrap: wrap;
gap: 10px;
margin-bottom: 30px;
justify-content: center;
}
.method-btn {
padding: 10px 20px;
border: 2px solid #667eea;
background: white;
color: #667eea;
border-radius: 25px;
cursor: pointer;
transition: all 0.3s;
font-weight: 600;
}
.method-btn:hover {
background: #667eea;
color: white;
transform: translateY(-2px);
}
.method-btn.active {
background: #667eea;
color: white;
}
.visualization-area {
background: #f8f9fa;
border-radius: 10px;
padding: 25px;
min-height: 400px;
}
h2 {
color: #333;
margin-bottom: 15px;
border-bottom: 3px solid #667eea;
padding-bottom: 10px;
}
.description {
color: #666;
margin-bottom: 20px;
line-height: 1.6;
font-style: italic;
}
.input-area {
margin-bottom: 20px;
}
input, textarea, select {
padding: 10px;
border: 2px solid #ddd;
border-radius: 5px;
font-size: 14px;
width: 100%;
margin-top: 5px;
}
button {
padding: 10px 25px;
background: #667eea;
color: white;
border: none;
border-radius: 5px;
cursor: pointer;
font-weight: 600;
margin-top: 10px;
}
button:hover {
background: #5568d3;
}
.output-area {
margin-top: 20px;
}
.table-container {
overflow-x: auto;
}
table {
width: 100%;
border-collapse: collapse;
margin-top: 15px;
background: white;
}
th, td {
padding: 12px;
text-align: left;
border: 1px solid #ddd;
}
th {
background: #667eea;
color: white;
font-weight: 600;
}
tr:nth-child(even) {
background: #f8f9fa;
}
.vector-display {
background: white;
padding: 15px;
border-radius: 5px;
margin: 10px 0;
border-left: 4px solid #667eea;
}
.vector {
font-family: 'Courier New', monospace;
color: #333;
margin: 5px 0;
}
.embedding-viz {
display: flex;
flex-wrap: wrap;
gap: 10px;
margin-top: 15px;
}
.embedding-item {
background: white;
padding: 15px;
border-radius: 8px;
border: 2px solid #667eea;
min-width: 200px;
}
.embedding-label {
font-weight: 600;
color: #667eea;
margin-bottom: 8px;
}
.embedding-vector {
font-family: 'Courier New', monospace;
font-size: 12px;
color: #666;
}
.text-processing {
background: white;
padding: 15px;
border-radius: 5px;
margin: 10px 0;
}
.word-item {
display: inline-block;
margin: 5px;
padding: 8px 15px;
background: #667eea;
color: white;
border-radius: 20px;
font-size: 14px;
}
.weight-display {
display: inline-block;
background: #764ba2;
padding: 3px 8px;
border-radius: 10px;
margin-left: 5px;
font-size: 12px;
}
label {
font-weight: 600;
color: #333;
display: block;
margin-top: 15px;
}
</style>
</head>
<body>
<div class="container">
<h1>🎓 Feature Encoding Methods Playground</h1>
<div class="method-selector">
<button class="method-btn active" onclick="showMethod('label')">Label Encoding</button>
<button class="method-btn" onclick="showMethod('onehot')">One-Hot Encoding</button>
<button class="method-btn" onclick="showMethod('embedding')">Embeddings</button>
<button class="method-btn" onclick="showMethod('highcard')">High Cardinality</button>
<button class="method-btn" onclick="showMethod('bow')">Bag-of-Words</button>
<button class="method-btn" onclick="showMethod('tfidf')">TF-IDF</button>
<button class="method-btn" onclick="showMethod('word2vec')">Word Embeddings</button>
<button class="method-btn" onclick="showMethod('pretrained')">Pre-trained</button>
<button class="method-btn" onclick="showMethod('transformer')">Transformers</button>
</div>
<div class="visualization-area" id="viz-area"></div>
</div>
<script>
const methods = {
label: {
title: "Label Encoding",
description: "Assigns integer IDs to categories. Best for ordinal features (e.g., Low < Medium < High).",
html: `
<div class="input-area">
<label>Enter categories (comma-separated):</label>
<input type="text" id="label-input" value="Low, Medium, High, Very High" placeholder="e.g., Small, Medium, Large">
<button onclick="applyLabelEncoding()">Encode</button>
</div>
<div class="output-area" id="label-output"></div>
`
},
onehot: {
title: "One-Hot Encoding (OHE)",
description: "Represents categories as binary vectors. Each category gets its own column with 1 or 0.",
html: `
<div class="input-area">
<label>Enter categories (comma-separated):</label>
<input type="text" id="onehot-input" value="Red, Blue, Green, Red, Blue" placeholder="e.g., Cat, Dog, Bird">
<button onclick="applyOneHot()">Encode</button>
</div>
<div class="output-area" id="onehot-output"></div>
`
},
embedding: {
title: "Embeddings",
description: "Dense vector representations learned during training. Captures relationships between categories in continuous space.",
html: `
<div class="input-area">
<label>Enter categories (comma-separated):</label>
<input type="text" id="embed-input" value="Dog, Cat, Tiger, Lion" placeholder="e.g., Apple, Banana, Orange">
<label>Embedding Dimension:</label>
<input type="number" id="embed-dim" value="3" min="2" max="10">
<button onclick="applyEmbedding()">Generate Embeddings</button>
</div>
<div class="output-area" id="embed-output"></div>
`
},
highcard: {
title: "High Cardinality Feature Handling",
description: "Maps thousands of categories (e.g., zip codes, product IDs) into compact vectors. Avoids sparse matrices from OHE.",
html: `
<div class="input-area">
<label>Simulate High Cardinality Feature (e.g., Product IDs):</label>
<input type="number" id="highcard-count" value="1000" min="100" max="10000" placeholder="Number of unique values">
<label>Target Embedding Dimension:</label>
<input type="number" id="highcard-dim" value="16" min="4" max="64">
<button onclick="applyHighCardinality()">Compare Approaches</button>
</div>
<div class="output-area" id="highcard-output"></div>
`
},
bow: {
title: "Bag-of-Words (BoW)",
description: "Represents text as raw word counts. Order is ignored, only frequency matters.",
html: `
<div class="input-area">
<label>Enter text documents (one per line):</label>
<textarea id="bow-input" rows="4">I love machine learning
Machine learning is amazing
Deep learning and machine learning</textarea>
<button onclick="applyBagOfWords()">Create BoW</button>
</div>
<div class="output-area" id="bow-output"></div>
`
},
tfidf: {
title: "TF-IDF",
description: "Weights words by importance: Term Frequency × Inverse Document Frequency. Rare words get higher weights.",
html: `
<div class="input-area">
<label>Enter text documents (one per line):</label>
<textarea id="tfidf-input" rows="4">The cat sat on the mat
The dog sat on the log
Cats and dogs are pets</textarea>
<button onclick="applyTFIDF()">Calculate TF-IDF</button>
</div>
<div class="output-area" id="tfidf-output"></div>
`
},
word2vec: {
title: "Word Embeddings (Word2Vec, GloVe, FastText)",
description: "Dense vector representations capturing semantic similarity. Similar words have similar vectors.",
html: `
<div class="input-area">
<label>Enter words to embed (comma-separated):</label>
<input type="text" id="word2vec-input" value="king, queen, man, woman, dog, cat" placeholder="e.g., happy, sad, joyful">
<label>Embedding Dimension:</label>
<input type="number" id="word2vec-dim" value="4" min="2" max="10">
<button onclick="applyWord2Vec()">Generate Word Embeddings</button>
</div>
<div class="output-area" id="word2vec-output"></div>
`
},
pretrained: {
title: "Pre-trained Embeddings",
description: "Load embeddings trained on large corpora (e.g., GloVe on Wikipedia). Can be frozen or fine-tuned.",
html: `
<div class="input-area">
<label>Select Pre-trained Model:</label>
<select id="pretrained-model">
<option value="glove">GloVe (Wikipedia)</option>
<option value="fasttext">FastText (Common Crawl)</option>
<option value="word2vec">Word2Vec (Google News)</option>
</select>
<label>Enter words to lookup:</label>
<input type="text" id="pretrained-input" value="artificial, intelligence, neural, network" placeholder="e.g., computer, science">
<button onclick="applyPretrained()">Lookup Embeddings</button>
</div>
<div class="output-area" id="pretrained-output"></div>
`
},
transformer: {
title: "Transformers (BERT, GPT)",
description: "Contextual embeddings using self-attention. Same word gets different embeddings based on context.",
html: `
<div class="input-area">
<label>Enter sentences to compare:</label>
<textarea id="transformer-input" rows="3">The bank by the river is steep
I need to go to the bank to deposit money</textarea>
<label>Select Model:</label>
<select id="transformer-model">
<option value="bert">BERT</option>
<option value="gpt">GPT</option>
</select>
<button onclick="applyTransformer()">Generate Contextual Embeddings</button>
</div>
<div class="output-area" id="transformer-output"></div>
`
}
};
function showMethod(method) {
document.querySelectorAll('.method-btn').forEach(btn => btn.classList.remove('active'));
event.target.classList.add('active');
const methodData = methods[method];
const vizArea = document.getElementById('viz-area');
vizArea.innerHTML = `
<h2>${methodData.title}</h2>
<p class="description">${methodData.description}</p>
${methodData.html}
`;
}
function applyLabelEncoding() {
const input = document.getElementById('label-input').value;
const categories = input.split(',').map(s => s.trim());
const uniqueCategories = [...new Set(categories)];
let html = '<div class="table-container"><table><tr><th>Category</th><th>Label</th></tr>';
uniqueCategories.forEach((cat, idx) => {
html += `<tr><td>${cat}</td><td>${idx}</td></tr>`;
});
html += '</table></div>';
html += '<div class="vector-display"><strong>Example Encoding:</strong><div class="vector">';
html += `Input: [${categories.join(', ')}]<br>`;
html += `Encoded: [${categories.map(c => uniqueCategories.indexOf(c)).join(', ')}]`;
html += '</div></div>';
document.getElementById('label-output').innerHTML = html;
}
function applyOneHot() {
const input = document.getElementById('onehot-input').value;
const categories = input.split(',').map(s => s.trim());
const uniqueCategories = [...new Set(categories)];
let html = '<div class="table-container"><table><tr><th>Original</th>';
uniqueCategories.forEach(cat => html += `<th>${cat}</th>`);
html += '</tr>';
categories.forEach(cat => {
html += `<tr><td><strong>${cat}</strong></td>`;
uniqueCategories.forEach(ucat => {
html += `<td>${cat === ucat ? '1' : '0'}</td>`;
});
html += '</tr>';
});
html += '</table></div>';
document.getElementById('onehot-output').innerHTML = html;
}
function applyEmbedding() {
const input = document.getElementById('embed-input').value;
const dim = parseInt(document.getElementById('embed-dim').value);
const categories = input.split(',').map(s => s.trim());
let html = '<div class="embedding-viz">';
categories.forEach((cat, idx) => {
const vector = Array(dim).fill(0).map(() => (Math.random() * 2 - 1).toFixed(3));
html += `
<div class="embedding-item">
<div class="embedding-label">${cat}</div>
<div class="embedding-vector">[${vector.join(', ')}]</div>
</div>
`;
});
html += '</div>';
html += '<p style="margin-top:15px;color:#666;"><em>Note: Vectors are randomly initialized. In real training, similar categories would have similar vectors.</em></p>';
document.getElementById('embed-output').innerHTML = html;
}
function applyHighCardinality() {
const count = parseInt(document.getElementById('highcard-count').value);
const dim = parseInt(document.getElementById('highcard-dim').value);
const oheSize = count;
const embSize = count * dim;
const compressionRatio = (oheSize / embSize).toFixed(2);
let html = `
<div class="vector-display">
<strong>Comparison for ${count} unique categories:</strong><br><br>
<strong>One-Hot Encoding:</strong><br>
• Matrix size: ${count} × ${count} = ${oheSize.toLocaleString()} values<br>
• Memory: Very sparse (mostly zeros)<br>
• Scalability: ❌ Poor for high cardinality<br><br>
<strong>Embedding Approach:</strong><br>
• Matrix size: ${count} × ${dim} = ${embSize.toLocaleString()} values<br>
• Memory: Dense, compact representation<br>
• Compression: ${compressionRatio}× smaller<br>
• Scalability: ✅ Excellent for high cardinality<br><br>
<strong>Example: Product ID "PROD_52847" → </strong>[${Array(dim).fill(0).map(() => (Math.random() * 2 - 1).toFixed(3)).join(', ')}]
</div>
`;
document.getElementById('highcard-output').innerHTML = html;
}
function applyBagOfWords() {
const input = document.getElementById('bow-input').value;
const docs = input.split('\n').filter(s => s.trim());
const allWords = new Set();
docs.forEach(doc => {
doc.toLowerCase().split(/\s+/).forEach(word => allWords.add(word));
});
const vocabulary = [...allWords].sort();
let html = '<div class="table-container"><table><tr><th>Document</th>';
vocabulary.forEach(word => html += `<th>${word}</th>`);
html += '</tr>';
docs.forEach((doc, idx) => {
const words = doc.toLowerCase().split(/\s+/);
const counts = {};
words.forEach(w => counts[w] = (counts[w] || 0) + 1);
html += `<tr><td><strong>Doc ${idx + 1}</strong></td>`;
vocabulary.forEach(word => {
html += `<td>${counts[word] || 0}</td>`;
});
html += '</tr>';
});
html += '</table></div>';
document.getElementById('bow-output').innerHTML = html;
}
function applyTFIDF() {
const input = document.getElementById('tfidf-input').value;
const docs = input.split('\n').filter(s => s.trim());
const allWords = new Set();
docs.forEach(doc => {
doc.toLowerCase().split(/\s+/).forEach(word => allWords.add(word));
});
const vocabulary = [...allWords].sort();
// Calculate IDF
const idf = {};
vocabulary.forEach(word => {
const docsWithWord = docs.filter(doc =>
doc.toLowerCase().includes(word)
).length;
idf[word] = Math.log(docs.length / docsWithWord);
});
let html = '<div class="text-processing"><strong>IDF Scores (Inverse Document Frequency):</strong><br>';
vocabulary.forEach(word => {
html += `<span class="word-item">${word}<span class="weight-display">${idf[word].toFixed(2)}</span></span>`;
});
html += '</div>';
html += '<div class="table-container"><table><tr><th>Document</th>';
vocabulary.forEach(word => html += `<th>${word}</th>`);
html += '</tr>';
docs.forEach((doc, idx) => {
const words = doc.toLowerCase().split(/\s+/);
const tf = {};
words.forEach(w => tf[w] = (tf[w] || 0) + 1);
Object.keys(tf).forEach(w => tf[w] /= words.length);
html += `<tr><td><strong>Doc ${idx + 1}</strong></td>`;
vocabulary.forEach(word => {
const tfidf = ((tf[word] || 0) * idf[word]).toFixed(3);
html += `<td>${tfidf}</td>`;
});
html += '</tr>';
});
html += '</table></div>';
document.getElementById('tfidf-output').innerHTML = html;
}
function applyWord2Vec() {
const input = document.getElementById('word2vec-input').value;
const dim = parseInt(document.getElementById('word2vec-dim').value);
const words = input.split(',').map(s => s.trim());
let html = '<div class="embedding-viz">';
words.forEach(word => {
const vector = Array(dim).fill(0).map(() => (Math.random() * 2 - 1).toFixed(3));
html += `
<div class="embedding-item">
<div class="embedding-label">${word}</div>
<div class="embedding-vector">[${vector.join(', ')}]</div>
</div>
`;
});
html += '</div>';
html += `
<div class="vector-display" style="margin-top:20px;">
<strong>Semantic Relationships (simulated):</strong><br>
• Similar words have similar vectors<br>
• Vector arithmetic: king - man + woman ≈ queen<br>
• Captures semantic meaning from context
</div>
`;
document.getElementById('word2vec-output').innerHTML = html;
}
function applyPretrained() {
const model = document.getElementById('pretrained-model').value;
const input = document.getElementById('pretrained-input').value;
const words = input.split(',').map(s => s.trim());
const modelInfo = {
glove: { name: 'GloVe', corpus: 'Wikipedia 2014 + Gigaword 5', dim: 300 },
fasttext: { name: 'FastText', corpus: 'Common Crawl (600B tokens)', dim: 300 },
word2vec: { name: 'Word2Vec', corpus: 'Google News (100B words)', dim: 300 }
};
const info = modelInfo[model];
let html = `
<div class="vector-display">
<strong>Model:</strong> ${info.name}<br>
<strong>Training Corpus:</strong> ${info.corpus}<br>
<strong>Dimension:</strong> ${info.dim}<br>
</div>
`;
html += '<div class="embedding-viz">';
words.forEach(word => {
const vector = Array(6).fill(0).map(() => (Math.random() * 2 - 1).toFixed(3));
html += `
<div class="embedding-item">
<div class="embedding-label">${word}</div>
<div class="embedding-vector">[${vector.join(', ')}, ...]</div>
<div style="font-size:11px;color:#999;margin-top:5px;">Showing 6/${info.dim} dims</div>
</div>
`;
});
html += '</div>';
html += `
<div class="vector-display" style="margin-top:20px;">
<strong>Usage:</strong><br>
• <strong>Frozen:</strong> Use pre-trained vectors as-is (transfer learning)<br>
• <strong>Fine-tuned:</strong> Update vectors during training on your task<br>
• Advantage: Leverage knowledge from massive corpora
</div>
`;
document.getElementById('pretrained-output').innerHTML = html;
}
function applyTransformer() {
const input = document.getElementById('transformer-input').value;
const model = document.getElementById('transformer-model').value;
const sentences = input.split('\n').filter(s => s.trim());
let html = `<div class="vector-display"><strong>Model: ${model.toUpperCase()}</strong> - Contextual embeddings using self-attention</div>`;
sentences.forEach((sent, idx) => {
const words = sent.split(/\s+/);
html += `<div class="text-processing"><strong>Sentence ${idx + 1}:</strong> "${sent}"<br><br>`;
words.forEach(word => {
const context_vector = Array(4).fill(0).map(() => (Math.random() * 2 - 1).toFixed(2));
html += `<span class="word-item">${word}<span class="weight-display">[${context_vector.join(', ')}]</span></span>`;
});
html += '</div>';
});
html += `
<div class="vector-display" style="margin-top:20px;">
<strong>Key Insight:</strong> The word "bank" gets DIFFERENT embeddings in each sentence:<br>
• Sentence 1 (river context): bank → [0.8, -0.3, 0.1, ...]<br>
• Sentence 2 (money context): bank → [-0.2, 0.9, 0.7, ...]<br><br>
<strong>Self-Attention:</strong> Each word attends to all other words to understand context
</div>
`;
document.getElementById('transformer-output').innerHTML = html;
}
// Initialize with Label Encoding
showMethod('label');
</script>
</body>
</html> |
