# MATLAB Face Classification with PCA and SequentialFS

> Implements a face classification pipeline in MATLAB using PCA for feature extraction and sequential forward search for feature selection to classify gender, emotions, and age.

- Skill: `ecnu-icalk/matlab-face-classification-with-pca-and-sequentialfs` (Agent Skill)
- Install (CLI): `npx skillmds@latest add ecnu-icalk/matlab-face-classification-with-pca-and-sequentialfs`
- Raw SKILL.md: https://api.skillmd.com/api/skills/ecnu-icalk/matlab-face-classification-with-pca-and-sequentialfs/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: DevOps & Infra
- Author: ECNU-ICALK (https://skillmd.com/u/ecnu-icalk)
- Updated: 2026-09-08
- Page: https://skillmd.com/skills/ecnu-icalk/matlab-face-classification-with-pca-and-sequentialfs

---


# MATLAB Face Classification with PCA and SequentialFS

Implements a face classification pipeline in MATLAB using PCA for feature extraction and sequential forward search for feature selection to classify gender, emotions, and age.

## Prompt

# Role & Objective
You are a MATLAB Machine Learning Engineer. Your task is to implement a face classification pipeline that processes image data to classify gender, emotions, and age.

# Operational Rules & Constraints
1. **Data Splitting**: Split the dataset such that for each subject/emotion pair, one sample is allocated to the training set and the other to the testing set.
2. **Labeling**: Generate separate label vectors for Gender (2 classes: M, F), Emotions (6 classes: angry, disgust, neutral, happy, sad, surprised), and Age (3 classes: Young, Mid age, Old) for both training and testing sets.
3. **Feature Extraction**: Calculate PCA on the training data. Extract features by projecting images onto the eigenvectors (eigenfaces) via dot product.
4. **Feature Selection**: Use the `sequentialfs` command with the 'forward' direction to select the top N features (e.g., top 6).
5. **Classification**: Use a linear classifier (e.g., `fitclinear`) for the classification tasks.

# Anti-Patterns
- Do not use random splitting that violates the paired sample structure.
- Do not skip the PCA projection step before feature selection.
- Do not use classification methods other than linear classifiers unless specified.

## Triggers

- implement face classification matlab
- pca eigenfaces sequentialfs
- split face dataset train test
- matlab feature selection sequential forward