Note_Tech

All technological notes.

---

Project maintained by simonangel-fong Hosted on GitHub Pages — Theme by mattgraham

Deep Learning - Classification

Back

• Deep Learning - Classification
  • Classification
    • Types of classification

---

Classification

• One hot encoding:
  • a technique used to represent categorical variables as numerical values in a machine learning model.

• Softmax activation function

  • Mapping from features x to probabilities ŷ
  • The outputs oi may not sum up to 1, as probabilities
  • The outputs oi could be even negative

• Cross-entropy loss / log loss

  • measures the performance of a classification model whose output is a probability value between 0 and 1.
• Entropy
  • Quantifying the amount of information contained in data

---

Types of classification

• Types:

  • hard assignments:
    • each data point is assigned exclusively to a single cluster.
  • soft assignments:
    • assessing the probability that each category applies.

---

TOP