Machine Learning Models

Tags: #◇Computer Science #◇CSINTSY #◇Machine Learning Models #◇Code Snippets #◇M4

---

◇Overview

Unlike traditional AI (symbolic AI) that uses a top-down approach by modeling rules to process data, Machine Learning employs a bottom-up approach. It feeds data and expected outputs into an algorithm to discover the underlying rules. A machine learning model acts as a hypothesis function that transforms input features $X$ into an output $y$.

Types of Machine Learning:

1. Supervised Learning: Models learn from labeled data to predict target variables.
   Classification: Predicts categorical variables (e.g., binary, multi-class).
   Regression: Predicts continuous/numerical variables.
2. Unsupervised Learning: Models identify patterns from data without labels (e.g., clustering).

Training Pipeline:
Data collection $\rightarrow$ Preprocessing $\rightarrow$ Train/Test split $\rightarrow$ Model training $\rightarrow$ Evaluation.
Models are evaluated on a test set to ensure they generalize well to unseen data.

◇Code Implementation

(No direct code provided, but the process involves dataset splitting, defining the model, and evaluation.)

◇Syntax Breakdown

(Not applicable)

◇Complexity Analysis

Evaluation Metrics:
Classification:
Accuracy: $\frac{TP + TN}{TP + TN + FP + FN}$
Precision: $\frac{TP}{TP + FP}$
Recall: $\frac{TP}{TP + FN}$
F1-Score: Harmonic mean of precision and recall.
Regression:
Mean Squared Error (MSE)
Root Mean Squared Error (RMSE)

Cross-Validation: K-Fold cross-validation divides the dataset into $K$ parts, using $K-1$ for training and $1$ for testing, repeating $K$ times and averaging the performance.

◇References & Navigation

• Previous Topic: ◇Adversarial Games
• Next Topic: ◇Decision Trees
• Related Notes: ◇Machine Learning, ◇Data Representation