MAALOUFimad02/Machine_Learning_Training
0
1<!DOCTYPE html>2<html lang="fr">3<head>4 <meta charset="UTF-8">5 <meta name="viewport" content="width=device-width, initial-scale=1.0">6 <title>Cours — ML Academy</title>7 <link rel="stylesheet" href="css/shared.css">8 <script>9 window.MathJax = {10 tex: { 11 inlineMath: [['$', '$'], ['\\(', '\\)']], 12 displayMath: [['$$', '$$'], ['\\[', '\\]']] 13 },14 options: { skipHtmlTags: ['script','noscript','style','textarea','pre'] }15 };16 </script>17 <script src="https://cdn.jsdelivr.net/npm/mathjax@3/es5/tex-mml-chtml.js"></script>18 <style>19 /* LAYOUT WITH SIDEBAR */20 .course-layout {21 display: flex;22 min-height: calc(100vh - var(--navbar-height));23 }24 25 /* SIDEBAR */26 .sidebar {27 width: var(--sidebar-width);28 background: var(--bg-secondary);29 border-right: 1px solid var(--border-color);30 position: fixed;31 top: var(--navbar-height);32 left: 0;33 bottom: 0;34 overflow-y: auto;35 z-index: 100;36 animation: fadeInLeft 0.5s ease;37 }38 39 .sidebar-header {40 padding: var(--space-xl);41 border-bottom: 1px solid var(--border-color);42 }43 44 .sidebar-badge {45 display: inline-block;46 padding: var(--space-xs) var(--space-sm);47 background: rgba(99, 102, 241, 0.1);48 border: 1px solid rgba(99, 102, 241, 0.3);49 border-radius: var(--radius-sm);50 font-family: 'JetBrains Mono', monospace;51 font-size: 0.65rem;52 color: var(--primary-light);53 text-transform: uppercase;54 letter-spacing: 1px;55 margin-bottom: var(--space-sm);56 }57 58 .sidebar-title {59 font-size: 1.1rem;60 font-weight: 700;61 color: var(--text-primary);62 margin-bottom: var(--space-xs);63 }64 65 .sidebar-subtitle {66 font-size: 0.8rem;67 color: var(--text-muted);68 }69 70 .sidebar-nav {71 padding: var(--space-md) 0;72 }73 74 .nav-section {75 margin-bottom: var(--space-md);76 }77 78 .nav-section-title {79 padding: var(--space-sm) var(--space-xl);80 font-family: 'JetBrains Mono', monospace;81 font-size: 0.65rem;82 color: var(--text-muted);83 text-transform: uppercase;84 letter-spacing: 1px;85 }86 87 .nav-item {88 display: flex;89 align-items: center;90 gap: var(--space-sm);91 padding: var(--space-sm) var(--space-xl);92 color: var(--text-secondary);93 text-decoration: none;94 font-size: 0.9rem;95 transition: all var(--transition-base);96 border-left: 3px solid transparent;97 }98 99 .nav-item:hover {100 background: var(--bg-hover);101 color: var(--text-primary);102 border-left-color: var(--primary);103 }104 105 .nav-item.active {106 background: rgba(99, 102, 241, 0.1);107 color: var(--primary-light);108 border-left-color: var(--primary);109 }110 111 .nav-icon {112 font-size: 1rem;113 font-family: 'JetBrains Mono', monospace;114 color: var(--text-muted);115 }116 117 .sidebar-footer {118 padding: var(--space-lg) var(--space-xl);119 border-top: 1px solid var(--border-color);120 margin-top: auto;121 }122 123 .progress-label {124 font-size: 0.75rem;125 color: var(--text-muted);126 margin-bottom: var(--space-sm);127 display: flex;128 justify-content: space-between;129 }130 131 .progress-bar {132 height: 6px;133 background: var(--bg-tertiary);134 border-radius: 3px;135 overflow: hidden;136 }137 138 .progress-fill {139 height: 100%;140 background: linear-gradient(90deg, var(--primary), var(--secondary));141 border-radius: 3px;142 transition: width 0.5s ease;143 animation: progress 1s ease;144 }145 146 /* PDF DOWNLOAD BUTTON */147 .pdf-download {148 display: flex;149 align-items: center;150 gap: var(--space-sm);151 padding: var(--space-sm) var(--space-md);152 background: rgba(239, 68, 68, 0.1);153 border: 1px solid rgba(239, 68, 68, 0.3);154 border-radius: var(--radius-md);155 color: #ef4444;156 font-size: 0.8rem;157 font-weight: 600;158 text-decoration: none;159 margin-top: var(--space-md);160 transition: all var(--transition-base);161 }162 163 .pdf-download:hover {164 background: rgba(239, 68, 68, 0.2);165 transform: translateY(-2px);166 }167 168 /* MAIN CONTENT */169 .main-content {170 flex: 1;171 margin-left: var(--sidebar-width);172 padding: var(--space-2xl);173 max-width: 900px;174 }175 176 /* COURSE HERO */177 .course-hero {178 text-align: center;179 padding: var(--space-2xl) 0;180 margin-bottom: var(--space-2xl);181 border-bottom: 1px solid var(--border-color);182 }183 184 .course-hero-badge {185 display: inline-flex;186 align-items: center;187 gap: var(--space-sm);188 padding: var(--space-xs) var(--space-md);189 background: rgba(16, 185, 129, 0.1);190 border: 1px solid rgba(16, 185, 129, 0.3);191 border-radius: 20px;192 font-size: 0.75rem;193 color: var(--success);194 margin-bottom: var(--space-lg);195 }196 197 .course-hero-badge .dot {198 width: 6px;199 height: 6px;200 background: var(--success);201 border-radius: 50%;202 animation: pulse 2s ease-in-out infinite;203 }204 205 .course-hero-title {206 font-size: clamp(1.75rem, 4vw, 2.5rem);207 font-weight: 700;208 color: var(--text-primary);209 margin-bottom: var(--space-md);210 }211 212 .course-hero-subtitle {213 font-size: 1rem;214 color: var(--text-secondary);215 margin-bottom: var(--space-xl);216 }217 218 .course-hero-stats {219 display: flex;220 justify-content: center;221 gap: var(--space-2xl);222 flex-wrap: wrap;223 }224 225 .course-hero-stat {226 text-align: center;227 }228 229 .course-hero-stat-value {230 font-family: 'JetBrains Mono', monospace;231 font-size: 2rem;232 font-weight: 700;233 color: var(--text-primary);234 }235 236 .course-hero-stat-label {237 font-size: 0.7rem;238 color: var(--text-muted);239 text-transform: uppercase;240 letter-spacing: 1px;241 margin-top: var(--space-xs);242 }243 244 /* SECTIONS */245 .section {246 margin-bottom: var(--space-3xl);247 scroll-margin-top: calc(var(--navbar-height) + var(--space-lg));248 }249 250 .section-header {251 margin-bottom: var(--space-xl);252 padding-bottom: var(--space-md);253 border-bottom: 1px solid var(--border-color);254 }255 256 .section-badge {257 display: inline-block;258 padding: var(--space-xs) var(--space-sm);259 background: rgba(99, 102, 241, 0.1);260 border-radius: var(--radius-sm);261 font-family: 'JetBrains Mono', monospace;262 font-size: 0.65rem;263 color: var(--primary-light);264 text-transform: uppercase;265 letter-spacing: 1px;266 margin-bottom: var(--space-sm);267 }268 269 .section-title {270 font-size: clamp(1.5rem, 3vw, 2rem);271 font-weight: 700;272 color: var(--text-primary);273 margin-bottom: var(--space-sm);274 }275 276 .section-title small {277 display: block;278 font-size: 0.9rem;279 color: var(--text-muted);280 font-weight: 500;281 margin-bottom: var(--space-xs);282 }283 284 .section-content p {285 font-size: 1rem;286 color: var(--text-secondary);287 line-height: 1.8;288 margin-bottom: var(--space-md);289 }290 291 .section-content strong {292 color: var(--text-primary);293 font-weight: 600;294 }295 296 /* CARDS GRID */297 .cards-grid {298 display: grid;299 grid-template-columns: repeat(auto-fit, minmax(250px, 1fr));300 gap: var(--space-md);301 margin: var(--space-lg) 0;302 }303 304 .info-card {305 background: var(--bg-card);306 border: 1px solid var(--border-color);307 border-radius: var(--radius-md);308 padding: var(--space-lg);309 transition: all var(--transition-base);310 }311 312 .info-card:hover {313 border-color: var(--primary);314 transform: translateY(-3px);315 }316 317 .info-card-icon {318 width: 40px;319 height: 40px;320 background: linear-gradient(135deg, var(--primary), var(--secondary));321 border-radius: var(--radius-md);322 display: flex;323 align-items: center;324 justify-content: center;325 font-size: 1rem;326 font-weight: 700;327 color: white;328 margin-bottom: var(--space-sm);329 }330 331 .info-card-title {332 font-size: 1rem;333 font-weight: 600;334 color: var(--text-primary);335 margin-bottom: var(--space-xs);336 }337 338 .info-card-text {339 font-size: 0.85rem;340 color: var(--text-secondary);341 }342 343 .info-card-tag {344 display: inline-block;345 margin-top: var(--space-sm);346 padding: var(--space-xs) var(--space-sm);347 background: var(--bg-tertiary);348 border-radius: var(--radius-sm);349 font-family: 'JetBrains Mono', monospace;350 font-size: 0.65rem;351 color: var(--text-muted);352 }353 354 /* PIPELINE */355 .pipeline {356 display: flex;357 flex-wrap: wrap;358 gap: var(--space-md);359 margin: var(--space-xl) 0;360 padding: var(--space-lg);361 background: var(--bg-card);362 border: 1px solid var(--border-color);363 border-radius: var(--radius-lg);364 }365 366 .pipeline-step {367 flex: 1;368 min-width: 120px;369 text-align: center;370 padding: var(--space-md);371 position: relative;372 }373 374 .pipeline-step:not(:last-child)::after {375 content: '->';376 position: absolute;377 right: -15px;378 top: 50%;379 transform: translateY(-50%);380 color: var(--primary);381 font-family: 'JetBrains Mono', monospace;382 font-size: 1rem;383 }384 385 .pipeline-num {386 width: 36px;387 height: 36px;388 background: linear-gradient(135deg, var(--primary), var(--secondary));389 border-radius: 50%;390 display: flex;391 align-items: center;392 justify-content: center;393 font-family: 'JetBrains Mono', monospace;394 font-size: 0.9rem;395 font-weight: 700;396 color: white;397 margin: 0 auto var(--space-sm);398 }399 400 .pipeline-label {401 font-size: 0.85rem;402 font-weight: 600;403 color: var(--text-primary);404 margin-bottom: var(--space-xs);405 }406 407 .pipeline-desc {408 font-size: 0.75rem;409 color: var(--text-muted);410 }411 412 /* CHECKLIST */413 .checklist {414 list-style: none;415 padding: 0;416 margin: var(--space-lg) 0;417 }418 419 .checklist li {420 display: flex;421 align-items: flex-start;422 gap: var(--space-sm);423 padding: var(--space-sm) 0;424 font-size: 0.95rem;425 color: var(--text-secondary);426 }427 428 .checklist li::before {429 content: '[v]';430 color: var(--success);431 font-family: 'JetBrains Mono', monospace;432 font-weight: 700;433 flex-shrink: 0;434 }435 436 /* EQUATIONS */437 .equation-block {438 background: var(--bg-card);439 border: 1px solid var(--border-color);440 border-left: 3px solid var(--primary);441 border-radius: var(--radius-md);442 padding: var(--space-lg);443 margin: var(--space-lg) 0;444 text-align: center;445 overflow-x: auto;446 }447 448 .equation-label {449 display: block;450 font-family: 'JetBrains Mono', monospace;451 font-size: 0.7rem;452 color: var(--text-muted);453 text-align: right;454 margin-top: var(--space-sm);455 }456 457 /* COMPARISON GRID */458 .comparison-grid {459 display: grid;460 grid-template-columns: repeat(2, 1fr);461 gap: var(--space-md);462 margin: var(--space-lg) 0;463 }464 465 .comparison-box {466 background: var(--bg-card);467 border: 1px solid var(--border-color);468 border-radius: var(--radius-md);469 padding: var(--space-lg);470 }471 472 .comparison-box h4 {473 font-size: 0.9rem;474 font-weight: 600;475 color: var(--text-primary);476 margin-bottom: var(--space-md);477 padding-bottom: var(--space-sm);478 border-bottom: 1px solid var(--border-color);479 }480 481 .comparison-box ul {482 list-style: none;483 padding: 0;484 margin: 0;485 }486 487 .comparison-box li {488 padding: var(--space-xs) 0;489 font-size: 0.85rem;490 color: var(--text-secondary);491 }492 493 .comparison-box li::before {494 content: '>';495 color: var(--primary);496 margin-right: var(--space-sm);497 font-family: 'JetBrains Mono', monospace;498 }499 500 /* METRIC CARDS */501 .metric-row {502 display: grid;503 grid-template-columns: repeat(auto-fit, minmax(200px, 1fr));504 gap: var(--space-md);505 margin: var(--space-lg) 0;506 }507 508 .metric-card {509 background: var(--bg-card);510 border: 1px solid var(--border-color);511 border-radius: var(--radius-md);512 padding: var(--space-lg);513 text-align: center;514 transition: all var(--transition-base);515 }516 517 .metric-card:hover {518 border-color: var(--primary);519 transform: translateY(-3px);520 }521 522 .metric-name {523 font-family: 'JetBrains Mono', monospace;524 font-size: 0.75rem;525 color: var(--text-muted);526 text-transform: uppercase;527 letter-spacing: 1px;528 margin-bottom: var(--space-sm);529 }530 531 .metric-formula {532 font-family: 'JetBrains Mono', monospace;533 font-size: 1rem;534 color: var(--primary-light);535 font-weight: 600;536 margin-bottom: var(--space-xs);537 }538 539 .metric-desc {540 font-size: 0.75rem;541 color: var(--text-muted);542 }543 544 /* RESPONSIVE */545 @media (max-width: 768px) {546 .sidebar {547 transform: translateX(-100%);548 transition: transform var(--transition-base);549 }550 551 .sidebar.open {552 transform: translateX(0);553 }554 555 .main-content {556 margin-left: 0;557 padding: var(--space-lg);558 }559 560 .comparison-grid {561 grid-template-columns: 1fr;562 }563 564 .pipeline-step:not(:last-child)::after {565 display: none;566 }567 }568 569 /* BACK TO TOP */570 .back-to-top {571 position: fixed;572 bottom: var(--space-xl);573 right: var(--space-xl);574 width: 44px;575 height: 44px;576 background: var(--primary);577 border: none;578 border-radius: 50%;579 color: white;580 font-size: 1.2rem;581 cursor: pointer;582 opacity: 0;583 visibility: hidden;584 transition: all var(--transition-base);585 z-index: 1000;586 display: flex;587 align-items: center;588 justify-content: center;589 }590 591 .back-to-top.visible {592 opacity: 1;593 visibility: visible;594 }595 596 .back-to-top:hover {597 background: var(--primary-light);598 transform: translateY(-3px);599 }600 601 /* AUTHOR FOOTER */602 .author-footer {603 background: var(--bg-secondary);604 border-top: 1px solid var(--border-color);605 padding: var(--space-xl);606 text-align: center;607 }608 609 .author-info {610 display: flex;611 justify-content: center;612 gap: var(--space-xl);613 flex-wrap: wrap;614 margin-bottom: var(--space-md);615 }616 617 .author-link {618 display: flex;619 align-items: center;620 gap: var(--space-sm);621 color: var(--text-secondary);622 text-decoration: none;623 font-size: 0.9rem;624 transition: color var(--transition-base);625 }626 627 .author-link:hover {628 color: var(--primary-light);629 }630 631 .author-link svg {632 width: 18px;633 height: 18px;634 fill: currentColor;635 }636 </style>637</head>638<body>639 <!-- Particles Background -->640 <div class="particles-container">641 <div class="particle"></div>642 <div class="particle"></div>643 <div class="particle"></div>644 <div class="particle"></div>645 <div class="particle"></div>646 </div>647 648 <!-- Navigation -->649 <nav class="navbar">650 <a href="index.html" class="navbar-brand">651 <div class="brand-logo">ML</div>652 <span>ML Academy</span>653 </a>654 <div class="navbar-nav">655 <a href="index.html" class="nav-link">656 <span class="nav-icon">[H]</span>657 <span>Accueil</span>658 </a>659 <a href="cours.html" class="nav-link active">660 <span class="nav-icon">[C]</span>661 <span>Cours</span>662 </a>663 <a href="tp.html" class="nav-link">664 <span class="nav-icon">[T]</span>665 <span>TPs</span>666 </a>667 <a href="feedback.html" class="nav-link">668 <span class="nav-icon">[F]</span>669 <span>Contact</span>670 </a>671 </div>672 <div class="nav-badge">673 <div class="dot"></div>674 <span>Google Colab Ready</span>675 </div>676 </nav>677 678 <!-- Course Layout -->679 <div class="course-layout">680 <!-- Sidebar -->681 <aside class="sidebar">682 <div class="sidebar-header">683 <div class="sidebar-badge">Formation 2025/2026</div>684 <h1 class="sidebar-title">Machine Learning</h1>685 <p class="sidebar-subtitle">Cours theoriques complets</p>686 <a href="cours.pdf" class="pdf-download" download>687 <svg width="16" height="16" viewBox="0 0 24 24" fill="currentColor">688 <path d="M19 9h-4V3H9v6H5l7 7 7-7zM5 18v2h14v-2H5z"/>689 </svg>690 Telecharger le PDF691 </a>692 </div>693 694 <nav class="sidebar-nav">695 <div class="nav-section">696 <div class="nav-section-title">Introduction</div>697 <a href="#intro" class="nav-item active">698 <span class="nav-icon">[1]</span>699 <span>Qu'est-ce que le ML ?</span>700 </a>701 </div>702 703 <div class="nav-section">704 <div class="nav-section-title">Algorithmes Supervises</div>705 <a href="#regression" class="nav-item">706 <span class="nav-icon">[2]</span>707 <span>Regression Lineaire</span>708 </a>709 <a href="#logistic" class="nav-item">710 <span class="nav-icon">[3]</span>711 <span>Regression Logistique</span>712 </a>713 <a href="#randomforest" class="nav-item">714 <span class="nav-icon">[4]</span>715 <span>Random Forest</span>716 </a>717 </div>718 719 <div class="nav-section">720 <div class="nav-section-title">Deep Learning</div>721 <a href="#neuralnets" class="nav-item">722 <span class="nav-icon">[5]</span>723 <span>Reseaux de Neurones</span>724 </a>725 <a href="#lstm" class="nav-item">726 <span class="nav-icon">[6]</span>727 <span>LSTM & Series Temp.</span>728 </a>729 </div>730 731 <div class="nav-section">732 <div class="nav-section-title">Evaluation</div>733 <a href="#metrics" class="nav-item">734 <span class="nav-icon">[7]</span>735 <span>Metriques</span>736 </a>737 <a href="#optimization" class="nav-item">738 <span class="nav-icon">[8]</span>739 <span>Optimisation</span>740 </a>741 </div>742 743 <div class="nav-section">744 <div class="nav-section-title">Navigation</div>745 <a href="index.html" class="nav-item">746 <span class="nav-icon">[H]</span>747 <span>Accueil</span>748 </a>749 <a href="tp.html" class="nav-item">750 <span class="nav-icon">[T]</span>751 <span>Travaux Pratiques</span>752 </a>753 <a href="feedback.html" class="nav-item">754 <span class="nav-icon">[F]</span>755 <span>Questions</span>756 </a>757 </div>758 </nav>759 760 <div class="sidebar-footer">761 <div class="progress-label">762 <span>Progression</span>763 <span id="progress-text">0%</span>764 </div>765 <div class="progress-bar">766 <div class="progress-fill" id="progress-fill" style="width: 0%"></div>767 </div>768 </div>769 </aside>770 771 <!-- Main Content -->772 <main class="main-content">773 <!-- Course Hero -->774 <div class="course-hero scroll-animate">775 <div class="course-hero-badge">776 <div class="dot"></div>777 <span>Pret a executer sur Google Colab</span>778 </div>779 <h1 class="course-hero-title">780 Cours de <span class="gradient-text">Machine Learning</span>781 </h1>782 <p class="course-hero-subtitle">783 Formation complete couvrant les algorithmes fondamentaux jusqu'aux techniques avancees.784 </p>785 <div class="course-hero-stats">786 <div class="course-hero-stat">787 <div class="course-hero-stat-value">8</div>788 <div class="course-hero-stat-label">Chapitres</div>789 </div>790 <div class="course-hero-stat">791 <div class="course-hero-stat-value">25+</div>792 <div class="course-hero-stat-label">Equations</div>793 </div>794 <div class="course-hero-stat">795 <div class="course-hero-stat-value">50+</div>796 <div class="course-hero-stat-label">Exemples de code</div>797 </div>798 </div>799 </div>800 801 <!-- Section 1: Introduction -->802 <section class="section" id="intro">803 <div class="section-header scroll-animate">804 <div class="section-badge">Partie 1 · 20 min</div>805 <h2 class="section-title">806 <small>Fondamentaux</small>807 Qu'est-ce que le Machine Learning ?808 </h2>809 </div>810 811 <div class="section-content scroll-animate">812 <p>813 Le <strong>Machine Learning (ML)</strong> est une branche de l'intelligence artificielle 814 qui permet aux machines d'apprendre a partir de donnees <em>sans etre explicitement programmees</em> 815 pour chaque tache. Au lieu de coder des regles a la main, on montre des exemples au modele 816 et il decouvre lui-meme les patterns.817 </p>818 819 <div class="callout callout-info scroll-animate">820 <div class="callout-icon">[i]</div>821 <div class="callout-content">822 <div class="callout-title">Idee fondamentale</div>823 <div class="callout-text">824 On cherche a approximer une fonction inconnue $f$ telle que $\hat{y} = f(x_1, x_2, \ldots, x_n)$. 825 Le modele ML apprend cette fonction a partir d'exemples $(x, y)$ connus.826 </div>827 </div>828 </div>829 830 <h3 style="font-size: 1.2rem; font-weight: 600; color: var(--text-primary); margin: var(--space-xl) 0 var(--space-md);">831 Types d'apprentissage832 </h3>833 834 <div class="cards-grid scroll-animate">835 <div class="info-card">836 <div class="info-card-icon">S</div>837 <h4 class="info-card-title">Supervise</h4>838 <p class="info-card-text">Donnees labellisees $(x, y)$ : regression et classification.</p>839 <span class="info-card-tag">Predictions</span>840 </div>841 <div class="info-card">842 <div class="info-card-icon">N</div>843 <h4 class="info-card-title">Non supervise</h4>844 <p class="info-card-text">Pas de labels : clustering, reduction de dimension.</p>845 <span class="info-card-tag">Patterns</span>846 </div>847 <div class="info-card">848 <div class="info-card-icon">R</div>849 <h4 class="info-card-title">Par renforcement</h4>850 <p class="info-card-text">Agent apprend via actions-recompenses.</p>851 <span class="info-card-tag">Strategies</span>852 </div>853 </div>854 855 <h3 style="font-size: 1.2rem; font-weight: 600; color: var(--text-primary); margin: var(--space-xl) 0 var(--space-md);">856 Pipeline ML typique857 </h3>858 859 <div class="pipeline scroll-animate">860 <div class="pipeline-step">861 <div class="pipeline-num">1</div>862 <div class="pipeline-label">Donnees</div>863 <div class="pipeline-desc">Collecte & nettoyage</div>864 </div>865 <div class="pipeline-step">866 <div class="pipeline-num">2</div>867 <div class="pipeline-label">Features</div>868 <div class="pipeline-desc">Engineering</div>869 </div>870 <div class="pipeline-step">871 <div class="pipeline-num">3</div>872 <div class="pipeline-label">Split</div>873 <div class="pipeline-desc">Train / Test</div>874 </div>875 <div class="pipeline-step">876 <div class="pipeline-num">4</div>877 <div class="pipeline-label">Modele</div>878 <div class="pipeline-desc">Entrainement</div>879 </div>880 <div class="pipeline-step">881 <div class="pipeline-num">5</div>882 <div class="pipeline-label">Evaluation</div>883 <div class="pipeline-desc">Metriques</div>884 </div>885 <div class="pipeline-step">886 <div class="pipeline-num">6</div>887 <div class="pipeline-label">Production</div>888 <div class="pipeline-desc">Deploiement</div>889 </div>890 </div>891 </div>892 </section>893 894 <!-- Section 2: Regression Lineaire -->895 <section class="section" id="regression">896 <div class="section-header scroll-animate">897 <div class="section-badge">Partie 2 · 25 min</div>898 <h2 class="section-title">899 <small>Algorithmes de base</small>900 Regression Lineaire901 </h2>902 </div>903 904 <div class="section-content scroll-animate">905 <p>906 La <strong>regression lineaire</strong> modelise la relation entre les features et la cible 907 par une fonction affine. C'est l'algorithme le plus simple mais souvent tres efficace 908 comme baseline.909 </p>910 911 <div class="equation-block">912 $$\hat{y} = w_0 + w_1 x_1 + w_2 x_2 + \cdots + w_n x_n = \mathbf{w}^T \mathbf{x}$$913 <span class="equation-label">Modele lineaire avec coefficients $\mathbf{w}$</span>914 </div>915 916 <p>917 L'objectif est de minimiser l'erreur quadratique moyenne (MSE) :918 </p>919 920 <div class="equation-block">921 $$\text{MSE} = \frac{1}{m} \sum_{i=1}^{m} (y_i - \hat{y}_i)^2$$922 <span class="equation-label">Fonction de cout : moyenne des erreurs quadratiques</span>923 </div>924 925 <div class="callout callout-info scroll-animate">926 <div class="callout-icon">[i]</div>927 <div class="callout-content">928 <div class="callout-title">Solution analytique</div>929 <div class="callout-text">930 La regression lineaire admet une solution fermee : 931 $\mathbf{w}^* = (X^T X)^{-1} X^T Y$. Pas besoin d'iterations !932 </div>933 </div>934 </div>935 936 <div class="comparison-grid scroll-animate">937 <div class="comparison-box">938 <h4>[+] Avantages</h4>939 <ul>940 <li>Tres rapide a entrainer</li>941 <li>Interpretable (coefficients)</li>942 <li>Pas d'hyperparametres</li>943 <li>Excellent baseline</li>944 </ul>945 </div>946 <div class="comparison-box">947 <h4>[-] Limitations</h4>948 <ul>949 <li>Relation lineaire uniquement</li>950 <li>Sensible aux outliers</li>951 <li>Performance decroit en haute dimension</li>952 </ul>953 </div>954 </div>955 </div>956 </section>957 958 <!-- Section 3: Regression Logistique -->959 <section class="section" id="logistic">960 <div class="section-header scroll-animate">961 <div class="section-badge">Partie 3 · 20 min</div>962 <h2 class="section-title">963 <small>Classification</small>964 Regression Logistique965 </h2>966 </div>967 968 <div class="section-content scroll-animate">969 <p>970 Malgre son nom, la <strong>regression logistique</strong> est un algorithme de <em>classification</em>. 971 Elle predit la probabilite d'appartenance a une classe en utilisant la fonction sigmoide.972 </p>973 974 <div class="equation-block">975 $$P(y=1|\mathbf{x}) = \sigma(\mathbf{w}^T \mathbf{x}) = \frac{1}{1 + e^{-\mathbf{w}^T \mathbf{x}}}$$976 <span class="equation-label">Fonction sigmoide pour la classification binaire</span>977 </div>978 979 <div class="callout callout-success scroll-animate">980 <div class="callout-icon">[v]</div>981 <div class="callout-content">982 <div class="callout-title">Cas d'usage : Dataset Titanic</div>983 <div class="callout-text">984 Predire la survie des passagers du Titanic a partir de leur age, sexe, 985 classe de billet, etc. Un classique du ML pour debuter !986 </div>987 </div>988 </div>989 </div>990 </section>991 992 <!-- Section 4: Random Forest -->993 <section class="section" id="randomforest">994 <div class="section-header scroll-animate">995 <div class="section-badge">Partie 4 · 30 min</div>996 <h2 class="section-title">997 <small>Ensemble Learning</small>998 Random Forest999 </h2>1000 </div>1001 1002 <div class="section-content scroll-animate">1003 <p>1004 <strong>Random Forest</strong> est un ensemble d'arbres de decision qui votent pour predire. 1005 C'est l'un des algorithmes les plus populaires en ML applique : performant, robuste, 1006 peu sensible au tuning.1007 </p>1008 1009 <h3 style="font-size: 1.2rem; font-weight: 600; color: var(--text-primary); margin: var(--space-xl) 0 var(--space-md);">1010 Algorithm : Bagging + Random Splits1011 </h3>1012 1013 <div class="pipeline scroll-animate">1014 <div class="pipeline-step">1015 <div class="pipeline-num">1</div>1016 <div class="pipeline-label">Bootstrap</div>1017 <div class="pipeline-desc">Echantillons aleatoires</div>1018 </div>1019 <div class="pipeline-step">1020 <div class="pipeline-num">2</div>1021 <div class="pipeline-label">Splits</div>1022 <div class="pipeline-desc">Features aleatoires</div>1023 </div>1024 <div class="pipeline-step">1025 <div class="pipeline-num">3</div>1026 <div class="pipeline-label">Arbres</div>1027 <div class="pipeline-desc">N arbres independants</div>1028 </div>1029 <div class="pipeline-step">1030 <div class="pipeline-num">4</div>1031 <div class="pipeline-label">Vote</div>1032 <div class="pipeline-desc">Moyenne ou mode</div>1033 </div>1034 </div>1035 1036 <div class="metric-row scroll-animate">1037 <div class="metric-card">1038 <div class="metric-name">n_estimators</div>1039 <div class="metric-formula">100 - 500</div>1040 <div class="metric-desc">Nombre d'arbres</div>1041 </div>1042 <div class="metric-card">1043 <div class="metric-name">max_depth</div>1044 <div class="metric-formula">10 - 30</div>1045 <div class="metric-desc">Profondeur max</div>1046 </div>1047 <div class="metric-card">1048 <div class="metric-name">min_samples_split</div>1049 <div class="metric-formula">2 - 10</div>1050 <div class="metric-desc">Min pour splitter</div>1051 </div>1052 </div>1053 1054 <div class="callout callout-success scroll-animate">1055 <div class="callout-icon">[*]</div>1056 <div class="callout-content">1057 <div class="callout-title">Feature Importance</div>1058 <div class="callout-text">1059 Random Forest fournit automatiquement l'importance de chaque feature, 1060 ce qui aide a comprendre quelles variables influencent le plus les predictions.1061 </div>1062 </div>1063 </div>1064 </div>1065 </section>1066 1067 <!-- Section 5: Neural Networks -->1068 <section class="section" id="neuralnets">1069 <div class="section-header scroll-animate">1070 <div class="section-badge">Partie 5 · 35 min</div>1071 <h2 class="section-title">1072 <small>Deep Learning</small>1073 Reseaux de Neurones1074 </h2>1075 </div>1076 1077 <div class="section-content scroll-animate">1078 <p>1079 Les <strong>reseaux de neurones</strong> sont inspires du cerveau humain : des couches de neurones 1080 interconnectes executent des transformations non-lineaires. Ils excellent pour les patterns complexes.1081 </p>1082 1083 <h3 style="font-size: 1.2rem; font-weight: 600; color: var(--text-primary); margin: var(--space-xl) 0 var(--space-md);">1084 Fonctionnement : Forward + Backprop1085 </h3>1086 1087 <div class="cards-grid scroll-animate">1088 <div class="info-card">1089 <div class="info-card-icon">F</div>1090 <h4 class="info-card-title">Forward Pass</h4>1091 <p class="info-card-text">Donnees traversent les couches : $\mathbf{h}_1 = \sigma(W_1 \mathbf{x} + b_1)$</p>1092 </div>1093 <div class="info-card">1094 <div class="info-card-icon">L</div>1095 <h4 class="info-card-title">Loss Computation</h4>1096 <p class="info-card-text">Compare prediction vs realite : $L = \frac{1}{m} \sum (y - \hat{y})^2$</p>1097 </div>1098 <div class="info-card">1099 <div class="info-card-icon">B</div>1100 <h4 class="info-card-title">Backpropagation</h4>1101 <p class="info-card-text">Calcule les gradients via la chaine de derivation</p>1102 </div>1103 <div class="info-card">1104 <div class="info-card-icon">G</div>1105 <h4 class="info-card-title">Gradient Descent</h4>1106 <p class="info-card-text">Met a jour les poids : $W \leftarrow W - \alpha \nabla_W L$</p>1107 </div>1108 </div>1109 1110 <h3 style="font-size: 1.2rem; font-weight: 600; color: var(--text-primary); margin: var(--space-xl) 0 var(--space-md);">1111 Fonctions d'activation1112 </h3>1113 1114 <div class="metric-row scroll-animate">1115 <div class="metric-card">1116 <div class="metric-name">ReLU</div>1117 <div class="metric-formula">$f(x) = \max(0, x)$</div>1118 <div class="metric-desc">Couches cachees</div>1119 </div>1120 <div class="metric-card">1121 <div class="metric-name">Sigmoid</div>1122 <div class="metric-formula">$f(x) = \frac{1}{1 + e^{-x}}$</div>1123 <div class="metric-desc">Classification binaire</div>1124 </div>1125 <div class="metric-card">1126 <div class="metric-name">Softmax</div>1127 <div class="metric-formula">$f(x_i) = \frac{e^{x_i}}{\sum_j e^{x_j}}$</div>1128 <div class="metric-desc">Classification multi-classe</div>1129 </div>1130 </div>1131 </div>1132 </section>1133 1134 <!-- Section 6: LSTM -->1135 <section class="section" id="lstm">1136 <div class="section-header scroll-animate">1137 <div class="section-badge">Partie 6 · 30 min</div>1138 <h2 class="section-title">1139 <small>Series Temporelles</small>1140 LSTM & Reseaux Recurrents1141 </h2>1142 </div>1143 1144 <div class="section-content scroll-animate">1145 <p>1146 <strong>LSTM (Long Short-Term Memory)</strong> est un type de reseau neuronal pour series temporelles. 1147 Il peut "retenir" l'information sur de longues periodes — crucial pour les predictions temporelles.1148 </p>1149 1150 <div class="callout callout-warning scroll-animate">1151 <div class="callout-icon">[!]</div>1152 <div class="callout-content">1153 <div class="callout-title">Probleme des RNN vanilla</div>1154 <div class="callout-text">1155 Les gradients disparaissent (vanishing) ou explosent (exploding) sur de longues sequences. 1156 Le LSTM resout ce probleme avec son <strong>cell state</strong>.1157 </div>1158 </div>1159 </div>1160 1161 <h3 style="font-size: 1.2rem; font-weight: 600; color: var(--text-primary); margin: var(--space-xl) 0 var(--space-md);">1162 Les trois portes du LSTM1163 </h3>1164 1165 <div class="metric-row scroll-animate">1166 <div class="metric-card">1167 <div class="metric-name">Forget Gate</div>1168 <div class="metric-formula">$f_t = \sigma(W_f [h_{t-1}, x_t] + b_f)$</div>1169 <div class="metric-desc">Quoi oublier ?</div>1170 </div>1171 <div class="metric-card">1172 <div class="metric-name">Input Gate</div>1173 <div class="metric-formula">$i_t = \sigma(W_i [h_{t-1}, x_t] + b_i)$</div>1174 <div class="metric-desc">Quoi ajouter ?</div>1175 </div>1176 <div class="metric-card">1177 <div class="metric-name">Output Gate</div>1178 <div class="metric-formula">$o_t = \sigma(W_o [h_{t-1}, x_t] + b_o)$</div>1179 <div class="metric-desc">Quoi exposer ?</div>1180 </div>1181 </div>1182 1183 <div class="callout callout-info scroll-animate">1184 <div class="callout-icon">[i]</div>1185 <div class="callout-content">1186 <div class="callout-title">Cas d'usage</div>1187 <div class="callout-text">1188 Prediction de prix boursiers, meteo, consommation energetique, 1189 traitement du langage naturel (NLP)...1190 </div>1191 </div>1192 </div>1193 </div>1194 </section>1195 1196 <!-- Section 7: Metriques -->1197 <section class="section" id="metrics">1198 <div class="section-header scroll-animate">1199 <div class="section-badge">Partie 7 · 25 min</div>1200 <h2 class="section-title">