Machine Learning

What do eigenvalues and eigenvectors mean in PCA?


Principal Component Analysis (PCA) is a powerful technique in the world of data analysis and machine learning, commonly used for dimensionality reduction and feature extraction. At the heart of PCA lie two essential concepts: eigenvalues and eigenvectors. In this blog post, we’ll delve into what these concepts mean in the context of PCA, and we’ll illustrate their significance with examples and images.

Understanding Eigenvectors

Eigenvectors are vectors associated with a square matrix, often the covariance matrix in PCA. When applied to PCA, eigenvectors take on a special role:

  1. Direction of Maximum Variance: Eigenvectors represent the directions in the original feature space along which the data varies the most. These directions are known as principal components.
  2. Orthogonality: Eigenvectors are not just any vectors; they are orthogonal to each other. In simple terms, this means that they are at right angles to one another. This orthogonality ensures that each eigenvector captures different aspects of the data’s variance, making them ideal for dimensionality reduction.
  3. Unit Vectors: Eigenvectors are unit vectors, meaning they have a length of 1. Their magnitude is not essential; their direction is what matters.
  4. Ranking by Eigenvalues: Eigenvectors are ranked by their associated eigenvalues in descending order. The eigenvector corresponding to the largest eigenvalue points in the direction of the maximum variance in the data, the second-largest eigenvalue corresponds to the second most significant direction, and so on.

Understanding Eigenvalues

Eigenvalues are scalar values associated with each eigenvector. In PCA, eigenvalues serve several critical purposes:

  1. Quantifying Variance: Eigenvalues quantify the amount of variance captured by the corresponding principal component (eigenvector). The larger the eigenvalue, the more variance that component explains.
  2. Ordering Significance: Eigenvalues are ordered in descending order, meaning the first eigenvalue corresponds to the most significant principal component, the second eigenvalue to the second most significant, and so forth.
  3. Proportion of Variance Explained: By dividing each eigenvalue by the sum of all eigenvalues, you can determine the proportion of the total variance in the data explained by each principal component. This helps in deciding how many principal components to retain for dimensionality reduction.

Example and Visualization

Let’s illustrate these concepts with a simple example in two dimensions (2D data).

Suppose we have the following data points:

Data Points:
(1, 2)
(2, 3)
(3, 4)
(4, 5)
(5, 6)

Let’s walk through a simple example with two features (2D data) to illustrate these concepts. Suppose we have a dataset with the following data points:

Data Points: (1, 2) (2, 3) (3, 4) (4, 5) (5, 6)

Center the data: First, we center the data by subtracting the mean from each data point. The mean of the data is (3, 4). After centering, our data becomes:

Centered Data: (-2, -2) (-1, -1) (0, 0) (1, 1) (2, 2)

Covariance matrix: Next, we compute the covariance matrix of the centered data. The covariance matrix is a symmetric matrix that tells us how features are correlated. In this case, the covariance matrix is:

Covariance Matrix: [[2.5 2.5 ] [2.5 2.5 ]]

Eigenvalues and Eigenvectors: Now, we calculate the eigenvalues and eigenvectors of the covariance matrix. In this simple example, both eigenvalues are the same (2.5), and the corresponding eigenvectors are

Eigenvector 1: [1, 1] 
Eigenvector 2: [-1, 1]

These eigenvectors represent the directions of maximum variance in our data. The first eigenvector represents the main axis of variation, and the second eigenvector represents the orthogonal direction.

Principal Components: The eigenvalues tell us the amount of variance captured by each principal component. Typically, you would retain the top-k principal components that explain the most variance while reducing the dimensionality of your data.

In summary, eigenvectors represent the directions of maximum variance (principal components), and eigenvalues quantify the amount of variance explained by each principal component. PCA utilizes these concepts to transform and reduce the dimensionality of data, facilitating data compression while preserving crucial information. Understanding eigenvalues and eigenvectors is fundamental to harnessing the power of PCA in data analysis and machine learning tasks.

Important Notice for college students

If you’re a college student and have skills in programming languages, Want to earn through blogging? Mail us at

For more Programming related blogs Visit Us Geekycodes . Follow us on Instagram.

Leave a Reply

%d bloggers like this: