Symmetric Matrices and Positive Definiteness

TL;DR

Positive definite matrices have various properties, including being invertible, the only positive definite projection matrix being the identity matrix, and diagonal matrices with positive entries also being positive definite. However, a symmetric matrix with a positive determinant may not necessarily be positive definite.

Transcript

PROFESSOR: Hi, everyone. Welcome back. So today, I'd like to talk about positive definite matrices. And specifically, we're going to analyze several properties of positive definite matrices. And specifically, we're going to look at why each one of these following statements is true. So first off, why every positive definite matrix is invertible. Wh... Read More

Key Insights

• 0️⃣ Every positive definite matrix is invertible because its determinant is non-zero, and all of its eigenvalues are greater than zero.
• 🟰 The only positive definite projection matrix is the identity matrix because it has eigenvalues equal to one.
• 🫤 Diagonal matrices with positive entries are positive definite because their product with any non-zero vector yields a positive value.
• 🫤 A symmetric matrix with a positive determinant may not be positive definite if it has negative values along its diagonal.
• ❓ The eigenvalue properties of matrices provide valuable insights into their positive definiteness.
• ❓ Proofs involving diagonalizable matrices and linear algebra techniques can establish properties of positive definite matrices.
• ❓ Counterexamples can be used to disprove generalizations about positive definite matrices.

Questions & Answers

Q: Why does a positive definite matrix need to have non-zero determinants?

A positive definite matrix has all eigenvalues greater than zero. Since the determinant of a matrix is the product of its eigenvalues, a non-zero determinant implies that none of the eigenvalues can be zero, making the matrix invertible.

Q: How can we prove that the identity matrix is the only positive definite projection matrix?

A projection matrix has eigenvalues only equal to zero or one. As a positive definite matrix must have all eigenvalues greater than zero, the only possible eigenvalue for a positive definite projection matrix is one. Thus, the identity matrix, which has all eigenvalues equal to one, is the only positive definite projection matrix.

Q: Why does a diagonal matrix with positive entries qualify as positive definite?

If we consider the product of a vector, x, and a diagonal matrix, D, using x^T * D * x, we obtain a sum of squares where the coefficients are the positive diagonal entries of D. As a positive number multiplied by a square (x_i^2) is positive, and the sum of positive numbers is positive, the product x^T * D * x is greater than zero, showing that D is positive definite.

Q: Can a symmetric matrix with a positive determinant be positive definite?

Not necessarily. To disprove this, we can construct a counterexample by choosing negative values along the diagonal of the symmetric matrix. By considering the product x^T * S * x, where x has a non-zero entry on the diagonal, we can show that the result is negative, indicating that the matrix is not positive definite.

Summary & Key Takeaways

• Positive definite matrices are invertible because their eigenvalues are all greater than zero, leading to a non-zero determinant.

• The only positive definite projection matrix is the identity matrix, as its eigenvalues are all equal to one.

• Diagonal matrices with positive entries are positive definite because the product of positive numbers is also positive.

• A symmetric matrix with a positive determinant may not be positive definite if it has negative entries along the diagonal.