🧠 Kuliah AI & Machine Learning

Belajar Artificial Intelligence & Machine Learning dari basic sampai advance, mencakup semua metode: Supervised → Unsupervised → Deep Learning → Ensemble → Reinforcement → NLP. Semua algoritma ditulis Python murni (tanpa library ML) dan bisa langsung dijalankan di simulator online — tanpa GPU, tanpa install apa pun.

Peta Kuliah AI & ML
Peta kuliah: 5 level, 18 bab, semua metode ML
💡 Bagaimana Simulator Ini Bekerja?

Kamu menulis kode Python — dan di sini kita belajar dengan cara menulis algoritmanya sendiri (gradient descent, k-means, backpropagation, Q-learning, dst.) supaya benar-benar paham "apa yang terjadi di dalam". Saat tombol Run Code ditekan, kode dijalankan di server dalam container Docker python:3.11-slim bersama ai_shim.py — generator data, statistik, metrik, dan visualisasi ASCII semuanya tersedia otomatis tanpa perlu import library ML.

Alur kerja simulator
Alur kerja: kode → server PHP → Docker Python + ai_shim.py → output

Fungsi helper yang disediakan (sudah otomatis tersedia di setiap editor):

# ── RNG deterministik ────────────────────────── random_seed(42) # agar hasil sama setiap Run (deterministik) rand() / randn() # angka acak seragam / normal randint(a, b) / choice(seq) / shuffle(lst) # ── Generator data sintetis ──────────────────── make_regression(n, slope, intercept, noise) # data regresi make_blobs(n, centers, spread) # data klaster (Gauss) make_moons(n) / make_circles(n) / make_xor(n) # data non-linear make_classification(n, n_features, sep) # data klasifikasi Gauss # ── Statistik & preprocessing ────────────────── mean() std() median() variance() correlation() normalize() # min-max ke [0,1] standardize() # z-score (mean 0, std 1) train_test_split(X, y, test_size) column_stack() add_bias() one_hot() # ── Linear algebra & aktivasi ────────────────── dot() matmul() transpose() identity() argmax() vec_add() vec_sub() vec_scale() sigmoid() relu() tanh_act() softmax() cross_entropy() euclidean() manhattan() cosine_sim() # ── Metrik evaluasi ──────────────────────────── accuracy() confusion_matrix() precision() recall() f1_score() log_loss() mse() rmse() mae() r2_score() # ── Visualisasi ASCII (diprint ke output) ────── plot_points(X, y, title) # scatter 2D, simbol per kelas scatter2d(points, title) # scatter manual [(x, y, 'char'), ...] line_chart(series, title) # grafik garis multi-seri cost_curve(losses) # kurva konvergensi histogram(xs, bins) # distribusi bar_chart(labels, values) # grafik batang heatmap(matrix, r, c) # matriks (confusion matrix) print_table(headers, ...)

⏱ Semua data deterministik (ber-seed): hasil Run sama persis setiap kali, jadi perubahan output pasti berasal dari perubahan logikamu. Program dihentikan otomatis setelah 8 detik — jaga agar training tidak terlalu panjang.

🟢 Level 1 — Dasar-Dasar

  • 1. Pengenalan AI & Machine Learning
    AI vs ML vs DL, 3 paradigma (supervised/unsupervised/RL), workflow ML, hello-world ML.
  • 2. Data, Statistik & Preprocessing
    Feature & label, statistik deskriptif, korelasi, normalisasi vs standardisasi.
  • 3. Evaluasi Model: Metrik & Train/Test Split
    Confusion matrix, accuracy, precision, recall, F1, MSE, RMSE, R², jebakan accuracy.

    • 🟡 Level 2 — Supervised Learning

  • 4. Linear Regression & Gradient Descent
    Model garis, cost function MSE, gradient descent, regresi berganda.
  • 5. Logistic Regression
    Sigmoid, probabilitas, cross-entropy, decision boundary.
  • 6. K-Nearest Neighbors (KNN)
    Jarak Euclidean, voting tetangga, efek nilai k, overfitting.
  • 7. Decision Tree
    Entropy, information gain, splitting, interpretasi, overfitting & depth.
  • 8. Naive Bayes
    Teorema Bayes, asumsi naive, smoothing, klasifikasi teks (spam).
  • 9. Support Vector Machine (SVM)
    Margin, support vector, hinge loss, regularisasi C.

    • 🟠 Level 3 — Unsupervised Learning

  • 10. K-Means Clustering
    Iterasi assign → update, inertia, elbow method, memilih k.
  • 11. Hierarchical Clustering & DBSCAN
    Agglomerative + linkage, dendrogram, DBSCAN, deteksi outlier.
  • 12. PCA: Dimensionality Reduction
    Varians, kovarians, komponen utama, power iteration, proyeksi.

    • 🔴 Level 4 — Neural Networks & Deep Learning

  • 13. Perceptron & Neural Network (Backpropagation)
    Perceptron, fungsi aktivasi, arsitektur MLP, forward & backward pass, XOR.
  • 14. CNN & RNN (Konsep + Simulasi)
    Konvolusi & pooling pada grid 2D, RNN Elman, hidden state, prediksi sekuens.

    • 🟣 Level 5 — Advanced

  • 15. Ensemble: Random Forest & Gradient Boosting
    Bagging, bootstrap, voting, boosting, residual, decision stump.
  • 16. Reinforcement Learning: Q-Learning
    Agent, state, action, reward, Q-table, Bellman update, ε-greedy, grid world.
  • 17. NLP: Teks Menjadi Angka
    Tokenisasi, bag-of-words, TF-IDF, cosine similarity, text classifier.
  • 18. Proyek Akhir (Capstone)
    Churn prediction, customer segmentation, fraud detection, end-to-end model comparison.
    • 📚 Prasyarat: dasar-dasar pemrograman Python (variabel, fungsi, loop, list, dictionary) dan sedikit aljabar (vektor, garis lurus). Jika belum, mulai dari Kuliah Pemrograman C (logika pemrograman), lalu lanjut ke sini.
    🎓 Kuliah lain di MyCourse: Pemrograman C • Microcontroller (Arduino) • Internet of Things