Electronics and

Telecommunication Engineering
FH2021
Engineering Mathematics IV

Mr. Ashwin Chavan

Assistant Professor
Department of Computer Engineering
SIES Graduate School of Technology

Mr. Ashwin Chavan | Quadratic forms

 Syllabus Content
Module: Linear Algebra: Quadratic Forms
5.1 Quadratic forms over real field, Linear Transformation of Quadratic form, Reduction of Quadratic form to
diagonal form using congruent transformation.
5.2 Rank, index, signature of quadratic form, Sylvester law of inertia, Value class of a quadratic form-Definite,
Semidefinite and Indefinite.
5.3 Reduction of Quadratic form to a canonical form using congruent transformations.
5.4 Singular Value Decomposition.
Self-learning Topics: Orthogonal Transformations, Applications of Quadratic forms and SVD in Engineering.

Course Outcome
At the end of the topic, student should be able to Reduce the Quadratic form to a canonical form using congruent and
orthogonal transformations

2
Mr. Ashwin Chavan | Quadratic forms
 Homogeneous polynomial of second degree in N variables is called a quadratic form

1. 𝑎𝑥 2 + 𝑏𝑦 2 + 2𝑐𝑥𝑦

2. Every quadratic form can be expressed in matrix notation as

𝑋𝐴𝑋 ′ , where X is a column matrix 𝑋′ is its transpose

3
Mr. Ashwin Chavan | Quadratic forms
 Homogeneous polynomial of second degree in N variables is called a quadratic form

1. Express 2𝑥 2 + 3𝑦 2 − 5𝑧 2 − 2𝑥𝑦 + 4𝑥𝑧 − 6𝑦𝑧 𝑖𝑛 𝑚𝑎𝑡𝑟𝑖𝑥 𝑓𝑜𝑟𝑚

4
Mr. Ashwin Chavan | Quadratic forms
 Homogeneous polynomial of second degree in N variables is called a quadratic form

1. Express 2𝑥12 − 3𝑥22 + 4𝑥32 + 𝑥42 − 2𝑥1 𝑥2 + 3𝑥1 𝑥3 − 4𝑥1 𝑥4 − 5𝑥2 𝑥3 + 6𝑥2 𝑥4 + 𝑥3 𝑥4 𝑖𝑛 𝑚𝑎𝑡𝑟𝑖𝑥 𝑓𝑜𝑟𝑚

5
Mr. Ashwin Chavan | Quadratic forms
 Linear Transformation of a Quadratic Form
Example: Express each of the following transformations
𝑥1 = 2𝑦1 − 3𝑦2 , 𝑥2 = 4𝑦1 + 𝑦2 𝑎𝑛𝑑 𝑦1 = 𝑧1 − 2𝑧2 , 𝑦2 = 2𝑧1 + 3𝑧2
In the form and find the composite transformation which expresses 𝑥1 𝑎𝑛𝑑 𝑥2 𝑖𝑛 𝑡𝑒𝑟𝑚𝑠 𝑜𝑓 𝑧1 𝑎𝑛𝑑 𝑧2

6
Mr. Ashwin Chavan | Quadratic forms
 Linear Transformation of a Quadratic Form

Consider the quadratic form 𝑋 ′ 𝐴𝑋, and the non singular transformation 𝑋 = 𝑃𝑌. [A is symmetric]

𝑋 ′ 𝐴𝑋 = 𝑃𝑌 ′ 𝐴𝑃𝑌 = 𝑌′ 𝑃′ 𝐴𝑃𝑌 = 𝑌′ 𝑃 ′ 𝐴𝑃 𝑌 = 𝑌′ 𝐵𝑌, 𝑤ℎ𝑒𝑟𝑒 𝑩 = 𝑷′ 𝑨𝑷

𝒀′ 𝑩𝒀 𝒊𝒔 𝒄𝒂𝒍𝒍𝒆𝒅 𝒍𝒊𝒏𝒆𝒂𝒓 𝒕𝒓𝒂𝒏𝒔𝒇𝒐𝒓𝒎 𝒐𝒇 𝒒𝒖𝒂𝒅𝒓𝒂𝒕𝒊𝒄 𝒇𝒐𝒓𝒎 𝑿′ 𝑨𝑿 𝒖𝒏𝒅𝒆𝒓 𝑿 = 𝑷𝒀

Congruence of a Square Matrix: A square matrix B of order n is said to be congruent to another square

matrix A of the same order, if there exists a non-singular matrix P such that 𝐵 = 𝑃’𝐴𝑃

7
Mr. Ashwin Chavan | Quadratic forms
Example 01: Reduce the following quadratic form to a diagonal form through congruent transformation ;
6𝑥12 + 3𝑥22 + 3𝑥32 − 4𝑥1 𝑥2 + 4𝑥1 𝑥3 − 2𝑥2 𝑥3 [Find diagonal matrix=P’AP]

8
Mr. Ashwin Chavan | Quadratic forms
 9
Mr. Ashwin Chavan | Quadratic forms
Example 02: Reduce the following quadratic form to a diagonal form through congruent transformation ;
6𝑥12 + 3𝑥22 + 14𝑥32 + 4𝑥1 𝑥2 + 18𝑥1 𝑥3 + 4𝑥2 𝑥3 [Find diagonal matrix=P’AP]

10
Mr. Ashwin Chavan | Quadratic forms
 11
Mr. Ashwin Chavan | Quadratic forms
 Rank, Index, Signature and Value class
• Rank (r): Number of non zero rows in B matrix is the rank of given quadratic form matrix
• Index (s): Number of positive square in 𝑌 ′𝐵𝑌 is called Index
• Signature: Let s be the number of positive squares then 𝑟 − 𝑠 will be the number of
negative squares. Then difference between positive and negative squares is called
signature of the quadratic form: 𝑆𝑖𝑔𝑛𝑎𝑟𝑢𝑟𝑒 = 𝑠 − 𝑟 − 𝑠 = 2𝑠 − 𝑟
• Value Class: 1) Positive definite: If all squared term of 𝑌 ′𝐵𝑌 are positive
2) Negative definite: If all squared term of 𝑌 ′𝐵𝑌 are negative
3) Positive semi definite: If all squared term of 𝑌 ′𝐵𝑌 are either positive or zero
4) Negative semi definite: If all squared term of 𝑌 ′𝐵𝑌 are either Negative or zero
5) Indefinite: Otherwise

Sylvester’s Law of Inertia: If s is a number of positive squares and r-s is number of

negative squares then law states that, “The signature of a real quadratic form is invariant ”
12
Mr. Ashwin Chavan | Quadratic forms
Ex.01: Reduce the following quadratic form 2𝑥12 + 𝑥22 − 3𝑥32 − 8𝑥2 𝑥3 − 4𝑥3 𝑥1 + 12𝑥1 𝑥2 to normal form
through congruent transformations. Also find its rank, signature and value class

13
Mr. Ashwin Chavan | Quadratic forms
 14
Mr. Ashwin Chavan | Quadratic forms
Ex.02: Reduce the following quadratic form 21𝑥12 + 11𝑥22 + 2𝑥32 − 8𝑥2 𝑥3 + 12𝑥3 𝑥1 − 30𝑥1 𝑥2 to normal
form / canonical form through congruent transformations. Also find its rank, signature, Also show that it is
positive semi definite. Find the non zero set of values of 𝑥1 , 𝑥2 𝑎𝑛𝑑 𝑥3 which will make quadratic form zero.

15
Mr. Ashwin Chavan | Quadratic forms
Ex.02: Reduce the following quadratic form 21𝑥12 + 11𝑥22 + 2𝑥32 − 8𝑥2 𝑥3 + 12𝑥3 𝑥1 − 30𝑥1 𝑥2 to normal
form / canonical form through congruent transformations. Also find its rank, signature, Also show that it is
positive semi definite. Find the non zero set of values of 𝑥1 , 𝑥2 𝑎𝑛𝑑 𝑥3 which will make quadratic form zero.

16
Mr. Ashwin Chavan | Quadratic forms
Ex.03: Reduce the following quadratic form 𝑥12 + 2𝑥22 + 3𝑥32 + 2𝑥2 𝑥3 − 2𝑥3 𝑥1 + 2𝑥1 𝑥2 to indefinite form.
Also find its rank, signature,

17
Mr. Ashwin Chavan | Quadratic forms
Ex.03: Reduce the following quadratic form 𝑥12 + 2𝑥22 + 3𝑥32 + 2𝑥2 𝑥3 − 2𝑥3 𝑥1 + 2𝑥1 𝑥2 to indefinite form.
Also find its rank, signature.

18
Mr. Ashwin Chavan | Quadratic forms
Ex.04: Reduce the following quadratic form 𝑥 2 − 2𝑦 2 + 10𝑧 2 − 10𝑥𝑦 + 4𝑥𝑧 − 2𝑦𝑧, to canonical form. Also
find its rank, signature, index and value class.

19
Mr. Ashwin Chavan | Quadratic forms
Ex.04: Reduce the following quadratic form 𝑥 2 − 2𝑦 2 + 10𝑧 2 − 10𝑥𝑦 + 4𝑥𝑧 − 2𝑦𝑧, to canonical form. Also
find its rank, signature, index and value class.

20
Mr. Ashwin Chavan | Quadratic forms
 Singular Value Decomposition [SVD]
1. 𝐹𝑖𝑛𝑑 𝐴’𝐴
2. Find Eigen values of 𝐴’𝐴
3. Arrange them in descending order, find their square roots and denote them by
𝜎1 𝑎𝑛𝑑 𝜎2 and find diagonal matrix D with them
4. Find Eigen vectors of 𝐴’𝐴, call them
𝑣1 𝑎𝑛𝑑 𝑣2 . 𝑁𝑜𝑡𝑒 𝑡ℎ𝑎𝑡 𝑡ℎ𝑒𝑦 𝑎𝑟𝑒 𝑜𝑟𝑡ℎ𝑜𝑔𝑜𝑛𝑎𝑙.
5. Normalise 𝑣1 𝑎𝑛𝑑 𝑣2 by dividing by their norms. And find the matrix 𝑉 =
[𝑣1 , 𝑣2 ]
1 1
6. To find 𝑈 = 𝑢1 , 𝑢2 , 𝑓𝑖𝑛𝑑 𝑢1 = 𝐴𝑣1 , 𝑢2 = 𝐴𝑣2
𝜎1 𝜎2
7. Then A can be expressed as 𝐴 = 𝑈𝐷𝑉′

21
Mr. Ashwin Chavan | Quadratic forms
 4 4
Ex.01: Find Singular value decomposition of 𝐴 =
−3 3

22
Mr. Ashwin Chavan | Quadratic forms
 4 4
Ex.01: Find Singular value decomposition of 𝐴 =
−3 3

23
Mr. Ashwin Chavan | Quadratic forms
 2 3
Ex.02: Find Singular value decomposition of 𝐴 =
0 2

24
Mr. Ashwin Chavan | Quadratic forms
 2 3
Ex.02: Find Singular value decomposition of 𝐴 =
0 2

25
Mr. Ashwin Chavan | Quadratic forms
 1 1
Ex.03: Find Singular value decomposition of 𝐴 = 1 1
1 −1

26
Mr. Ashwin Chavan | Quadratic forms
 1 1
Ex.03: Find Singular value decomposition of 𝐴 = 1 1
1 −1

27
Mr. Ashwin Chavan | Quadratic forms