MSK Institute • Shikohabad
Loading Page Content...
Preparing courses, live labs, and learning resources.सामग्री लोड हो रही है, कृपया प्रतीक्षा करें...
Connecting serverPlease wait...
💻 Practical Labs🚀 Free Demo📜 Verified Certificates
Matrix Determinant, Inversion & Solving Linear Systems0%
Matrix Determinant, Inversion & Solving Linear Systems
Beginner12 min read•Updated: Jul 11, 2026
Solving Linear Systems: inv(), det(), & solve()
Systems of simultaneous linear equations arise everywhere: in engineering simulations, structural load calculations, econometric models, and machine learning normal equations.
A system of linear equations with $n$ variables is represented in matrix notation as:
$$A \mathbf{x} = \mathbf{b}$$
Where:
- $A$ is the $n \times n$ coefficient matrix.
- $\mathbf{x}$ is the vector of unknown variables.
- $\mathbf{b}$ is the vector of constants.
NumPy's linear algebra module (numpy.linalg) provides tools to solve these systems with numerical stability.
Click to Zoom1. Matrix Determinants: np.linalg.det()
A square matrix $A$ has a unique inverse if and only if its determinant is non-zero ($\det(A) \neq 0$). A matrix with $\det(A) = 0$ is called singular or non-invertible:
import numpy as np
A = np.array([[4.0, 7.0],
[2.0, 6.0]])
# Compute determinant: (4*6) - (7*2) = 24 - 14 = 10
determinant = np.linalg.det(A)
print(f"Determinant: {determinant:.2f}") # 10.00
2. Solving Systems: Why You Should NEVER Use inv()!
Mathematically, the analytical solution to $A \mathbf{x} = \mathbf{b}$ is:
$$\mathbf{x} = A^{-1} \mathbf{b}$$
Many beginners compute this directly using np.linalg.inv(A) @ b.
Important Warning
Never invert a matrix to solve a linear system!
Direct matrix inversion is:
1. Numerically Unstable: Inversion magnifies floating-point roundoff errors and condition number instabilities.
2. Computationally Inefficient: Computing $A^{-1}$ takes $\mathcal{O}(n^3)$ operations with high constant factors.
Always use
np.linalg.solve(A, b)! solve() uses LAPACK's optimized LU decomposition with partial pivoting, which is significantly faster and far more numerically precise.# Solve the linear system:
# 3x + 1y = 9
# 1x + 2y = 8
A = np.array([[3.0, 1.0],
[1.0, 2.0]])
b = np.array([9.0, 8.0])
# The PROFESSIONAL way:
x = np.linalg.solve(A, b)
print("Solution x, y:", x) # [2., 3.] -> 3*(2) + 1*(3) = 9; 1*(2) + 2*(3) = 8!
# Verify solution by computing residual norm: ||Ax - b||
residual = np.linalg.norm(A @ x - b)
print(f"Residual error: {residual:.2e}") # 0.00e+00 (exact!)
3. Matrix Inversion: np.linalg.inv()
When theoretical requirements strictly dictate computing the explicit inverse matrix $A^{-1}$ (such as calculating covariance parameter matrices in statistics):
A = np.array([[1.0, 2.0],
[3.0, 4.0]])
A_inv = np.linalg.inv(A)
print("Inverse matrix A^-1:
", A_inv)
# Verify that A @ A^-1 is the identity matrix I:
identity_check = A @ A_inv
print("
A @ A^-1:
", np.round(identity_check, 4))
Output:
Inverse matrix A^-1:
[[-2. 1. ]
[ 1.5 -0.5]]
A @ A^-1:
[[1. 0.]
[0. 1.]]
4. Singular Matrices and Least-Squares: np.linalg.lstsq()
If a matrix is non-square (more equations than unknowns) or singular ($det(A) = 0$), np.linalg.solve() raises LinAlgError: Singular matrix.
To find the best-fit approximate solution that minimizes the sum of squared residuals $\|A\mathbf{x} - \mathbf{b}\|^2_2$, use Linear Least Squares (np.linalg.lstsq):
# Fitting a linear trend y = mx + c to noisy data
# 4 data points (x, y): (0, 1), (1, 2.2), (2, 2.9), (3, 4.1)
x_pts = np.array([0, 1, 2, 3])
y_pts = np.array([1.0, 2.2, 2.9, 4.1])
# Construct design matrix [x, 1]
A_design = np.column_stack((x_pts, np.ones_like(x_pts)))
# Solve least-squares:
params, residuals, rank, s = np.linalg.lstsq(A_design, y_pts, rcond=None)
slope, intercept = params
print(f"Fitted Line: y = {slope:.2f}x + {intercept:.2f}")
Interactive Quiz
Multiple Choice Questions
Question 1
What does a matrix determinant equal to zero ($det(A) = 0$) indicate?
Question 2
Why is
np.linalg.solve(A, b) preferred over np.linalg.inv(A) @ b?Question 3
Which function solves overdetermined or singular linear systems by minimizing squared errors?
Question 4
What is the expected result of multiplying an invertible matrix $A$ with its inverse $A^{-1}$ ($A @ A^{-1}$)?
Question 5
What exception is raised if
np.linalg.solve() is given a singular matrix?Next Lesson
Eigenvalues, Eigenvectors & Singular Value Decomposition (SVD)
Continue learning with hands-on practice, examples, and exercises in the upcoming topic.
Related Lessons
Practice Quiz
Test your understanding of this lesson with 5 questions. Each question has one correct answer.