Principal-Component-Analysis
Principal Component Analysis (PCA) is a technique which is used for dimension reduction.
Purpose
- Dimension reduction, keep the most important dimenstions
- Find principle component
- Project data a new space
Develop tools and techniques
- Python
- Pycharm
Skill
- Linear algebra: eigen decomposition
- Vector projection
Implementation Step
- Create data matrix which contains a lot of row vectors
- Normalize the matrix, which transforms all dimensions to the distribution of zero mean
- Calculate the covariance matrix,
- Do eigen decomposition and get the eigen values and eigen vectors
- Verify whether the decomposition is right or not,
- Project original data to the new space by the eigen vectors,
- Visualization
Result
