1 Definition and notation
A matrix is a rectangular arrangement of entries, usually written in rows and columns. The entries may be numbers, variables, symbols, or more general expressions. In elementary algebra and linear algebra, matrices are used to represent structured data and to encode operations that would otherwise require lengthy lists of equations or calculations.
Matrix notation provides a compact language for describing linear relationships. A matrix may be denoted by a capital letter such as A, B, or M, while its individual entries are referenced with subscripts. This framework supports both abstract theory and practical computation.
1.1 Basic structure
A matrix consists of horizontal rows and vertical columns. The entry in the i-th row and j-th column is often written as aij. The arrangement is ordered, so changing the position of an entry changes the matrix.
Matrices may contain repeated values, symbolic expressions, or parameters. In many applications, each row represents one object or observation, while each column represents one feature or variable. This structure makes matrices useful for organizing information in a regular format.
1.2 Order and dimensions
The size of a matrix is described by its order or dimensions, written as m × n, meaning m rows and n columns. A matrix with m rows and n columns is called an m-by-n matrix. The number of entries is therefore mn.
The order determines which operations are allowed. For example, matrices can be added only when they have the same dimensions, while multiplication requires a compatible number of columns and rows. These constraints are central to matrix algebra.
1.3 Element notation
Entries are commonly indexed using subscripts. For a matrix A, the element in row i and column j is denoted aij. This notation is standard in mathematical writing and is especially useful when referring to specific locations within a matrix.
Alternative notations also appear in practice, including bracketed arrays and parenthesized forms. In contexts involving vectors or operators, notation may be adapted to emphasize structure rather than individual entries. Regardless of style, the indexing convention remains the same.
1.4 Special types of matrices
Certain matrices occur so often that they are given special names. These types usually have distinctive patterns in their entries and often simplify calculations.
1.4.1 Square matrix
A square matrix has the same number of rows and columns. Such matrices are central in linear algebra because many important concepts, including determinants, eigenvalues, and inverses, are defined primarily for them.
Square matrices often represent linear transformations from a space to itself. Their algebraic properties are more extensive than those of nonsquare matrices, and they form the setting for many major theorems.
1.4.1.1 Diagonal matrix
A diagonal matrix is a square matrix whose nonzero entries, if any, appear only on the main diagonal. All entries away from the diagonal are zero. This form makes many operations especially simple.
Diagonal matrices are easy to multiply, invert when the diagonal entries are nonzero, and analyze in terms of their diagonal values. They frequently arise in simplified models and in matrix decompositions.
1.4.1.2 Identity matrix
The identity matrix has 1s on the main diagonal and 0s elsewhere. It acts as the multiplicative identity for square matrices of the same size, meaning that multiplying a matrix by the identity leaves it unchanged.
Identity matrices play a role analogous to the number 1 in ordinary arithmetic. They are used in defining inverses, constructing algorithms, and expressing linear transformations that do not alter vectors.
1.4.2 Zero matrix
The zero matrix has every entry equal to 0. It serves as the additive identity, since adding it to a matrix leaves the matrix unchanged. Zero matrices may have any dimension.
Because every entry is zero, this matrix represents the absence of a linear effect in many settings. It is also used as a reference point in proofs and computations.
1.4.3 Row matrix and column matrix
A row matrix has a single row and multiple columns, while a column matrix has a single column and multiple rows. These are sometimes treated as special cases of vectors written in matrix form.
Row and column matrices are commonly used in representing coordinates, linear functionals, and data lists. Their shape makes them convenient for computations involving transposes and dot products.
1.4.4 Symmetric and skew-symmetric matrices
A symmetric matrix is a square matrix equal to its transpose, so its entries mirror across the main diagonal. Symmetry often reflects balance in a system or correspondence between paired interactions.
A skew-symmetric matrix satisfies the condition that its transpose is the negative of the matrix itself. In such matrices, diagonal entries must be zero. These forms are important in geometry, mechanics, and theoretical linear algebra.
2 Matrix operations
Matrices support several fundamental operations. These operations generalize familiar arithmetic and are designed to preserve the rectangular structure of the data. Some are defined entrywise, while others depend on the interaction between rows and columns.
2.1 Addition and subtraction
Matrices of the same order can be added or subtracted by combining corresponding entries. If two matrices have matching dimensions, their sum or difference is formed element by element.
These operations are straightforward and obey familiar rules such as commutativity and associativity. They are used to combine data sets, compare matrices, and build more complex expressions.
2.2 Scalar multiplication
Scalar multiplication means multiplying every entry of a matrix by the same number or scalar. This changes the magnitude of the matrix without altering its shape.
The operation distributes across matrix addition and interacts naturally with other algebraic rules. In geometric interpretations, scalar multiplication can enlarge, shrink, or reverse the effect of a matrix.
2.3 Matrix multiplication
Matrix multiplication combines matrices in a way that reflects composition of linear transformations and interaction among rows and columns. Unlike entrywise multiplication, it is defined through sums of products.
The product of two matrices is generally formed by taking dot products of rows of the first matrix with columns of the second. This operation is fundamental in linear algebra and widely used in computation.
2.3.1 Rules and compatibility
To multiply two matrices, the number of columns in the first matrix must equal the number of rows in the second. If A is m × n and B is n × p, then the product AB is defined and has order m × p.
This compatibility condition ensures that each entry in the product can be computed from a row-column pairing. If the dimensions do not match, multiplication is not defined.
2.3.2 Non-commutativity
Matrix multiplication is usually not commutative, meaning AB often differs from BA even when both products exist. This contrasts with ordinary multiplication of numbers.
The order of multiplication matters because it reflects the order in which transformations are applied. In many applications, reversing the order changes the outcome in a substantial way.
2.4 Transpose
The transpose of a matrix is obtained by switching rows and columns. The first row becomes the first column, the second row becomes the second column, and so on.
Transpose operations are useful for expressing symmetry, defining dot products, and forming adjoints in more advanced settings. They also appear frequently in data analysis and optimization.
2.5 Determinant
The determinant is a scalar value associated with a square matrix. It summarizes certain algebraic and geometric properties of the matrix, including whether it is invertible and how it scales area or volume under a transformation.
A nonzero determinant indicates that the matrix has full rank in the square case and usually corresponds to a reversible linear transformation. A zero determinant signals degeneracy, such as flattening or collapse along some direction.
2.6 Inverse matrix
An inverse matrix, when it exists, is a matrix that reverses the effect of another square matrix under multiplication. If A has inverse A−1, then AA−1 and A−1A both equal the identity matrix.
Not every square matrix has an inverse. Matrices with zero determinant do not, and such matrices are called singular. When an inverse exists, it is central to solving linear systems and manipulating linear transformations.
3 Matrix algebra and theory
Matrix theory studies structural properties that go beyond individual computations. These concepts help classify matrices and explain how they behave under algebraic operations and transformations.
3.1 Rank
The rank of a matrix measures the dimension of the space spanned by its rows or columns. It indicates how much independent information the matrix contains.
Rank is closely connected to solvability of linear systems, invertibility, and dimensional reduction. A matrix with high rank preserves more information than one with low rank.
3.2 Trace
The trace of a square matrix is the sum of its diagonal entries. It is a simple invariant with important algebraic properties.
Trace appears in formulas involving eigenvalues, similarity transformations, and matrix functions. In applications, it can summarize certain cumulative effects of a transformation or system.
3.3 Eigenvalues and eigenvectors
An eigenvector of a square matrix is a nonzero vector whose direction is preserved under multiplication by the matrix, up to a scalar factor. That scalar is called the corresponding eigenvalue.
These concepts reveal the internal action of a matrix on a vector space. They are essential in differential equations, stability analysis, vibration theory, and many other areas.
3.4 Characteristic polynomial
The characteristic polynomial of a square matrix is formed from the determinant of a matrix expression involving a variable and the identity matrix. Its roots are the eigenvalues of the matrix.
This polynomial encodes key information about the matrix in algebraic form. It is used to study spectral properties and to derive identities related to powers of matrices.
3.5 Similarity and equivalence
Two matrices are similar if one can be obtained from the other by a change of basis using an invertible matrix. Similar matrices represent the same linear transformation in different coordinates.
Equivalence is a broader relation that often involves multiplication by invertible matrices on the left and right. These relations help classify matrices according to structural features rather than superficial entry patterns.
4 Matrices in linear systems
Matrices provide a standard way to represent systems of linear equations. By organizing coefficients into rows and columns, they make it easier to study and solve collections of equations systematically.
4.1 Systems of linear equations
A linear system consists of several equations in several unknowns, each equation being linear in those variables. The coefficients of the variables can be arranged into a matrix.
Matrix methods make it possible to analyze the system as a single object. This approach is especially useful when the system has many equations or when a direct substitution method would be cumbersome.
4.2 Augmented matrices
An augmented matrix combines the coefficient matrix of a system with the column of constants on the right-hand side. It provides a compact representation of the entire system.
This format is convenient for row operations and elimination methods. It allows the system to be manipulated without repeatedly rewriting the equations themselves.
4.3 Row reduction
Row reduction is the process of applying elementary row operations to simplify a matrix. The goal is often to reach row-echelon form or reduced row-echelon form.
These simplified forms reveal structural information such as pivots, free variables, and consistency. Row reduction is a standard method for solving linear systems and finding matrix rank.
4.4 Gaussian elimination
Gaussian elimination is a systematic row-reduction procedure used to solve linear systems. It eliminates variables step by step, creating an upper-triangular or echelon form.
The method is one of the most widely used algorithms in linear algebra. It is efficient, general, and forms the basis for many computational techniques in numerical mathematics.
4.5 Existence and uniqueness of solutions
A linear system may have no solution, exactly one solution, or infinitely many solutions. Matrix methods help determine which case occurs by examining rank and row-reduced form.
Uniqueness typically depends on whether the coefficient matrix has full rank in the appropriate setting. When the system is consistent but underdetermined, free variables lead to multiple solutions.
5 Matrix decompositions
Matrix decomposition breaks a matrix into factors with simpler structure. These factorizations are valuable for computation, proof, and interpretation.
5.1 LU decomposition
LU decomposition factors a matrix into a lower triangular matrix and an upper triangular matrix. This form is especially useful for solving linear systems efficiently after the factorization has been found.
The method reduces repeated work when the same coefficient matrix is used with multiple right-hand sides. It also appears in numerical algorithms and matrix analysis.
5.2 QR decomposition
QR decomposition writes a matrix as the product of an orthogonal or unitary factor and an upper triangular factor. It is widely used in least-squares problems and numerical stability analysis.
Because orthogonal transformations preserve lengths and angles, QR decomposition often yields robust computational procedures. It is also important in eigenvalue algorithms.
5.3 Cholesky decomposition
Cholesky decomposition factors a symmetric positive definite matrix into the product of a lower triangular matrix and its transpose. It is more specialized than LU decomposition but can be faster and more stable in its domain.
This decomposition is common in optimization, statistics, and simulation. Its existence depends on strong structural conditions on the matrix.
5.4 Singular value decomposition
Singular value decomposition, often abbreviated SVD, expresses a matrix as a product of two orthogonal or unitary matrices and a diagonal matrix of singular values. It applies to general matrices, not only square ones.
SVD is one of the most powerful tools in matrix theory. It supports data compression, noise reduction, rank approximation, and the analysis of linear maps in a highly structured way.
6 Applications
Matrices are used across science and technology because they organize complex relationships in a form suited to calculation and analysis. Their flexibility makes them useful in both theoretical and applied settings.
6.1 Geometry and transformations
In geometry, matrices represent rotations, reflections, scaling, shearing, and other linear transformations. They describe how points and vectors move under these operations.
This representation simplifies geometric reasoning by translating spatial changes into algebraic rules. It also makes composition of transformations straightforward through matrix multiplication.
6.2 Physics and engineering
Matrices are used to model mechanical systems, electrical circuits, control systems, and other physical processes. They can represent forces, state transitions, stress relations, and coupled equations.
In engineering, matrix methods support design, simulation, and stability analysis. Their compact form helps manage large networks of interacting components.
6.3 Computer graphics
Computer graphics relies heavily on matrices to transform objects in two and three dimensions. Translation, rotation, scaling, and perspective-related computations are often implemented using matrix operations.
These techniques allow efficient manipulation of scenes, animations, and camera views. Matrix pipelines are a standard part of rendering systems.
6.4 Statistics and data analysis
Matrices organize data tables, covariance structures, regression models, and multivariate observations. They are essential in methods that handle many variables at once.
In statistics, matrix formulations simplify estimation and inference. They also underpin principal component analysis, least squares fitting, and many machine learning methods.
6.5 Network analysis
Matrices can encode connections in graphs and networks, such as links between nodes or weights on edges. An adjacency matrix is a common example.
This representation supports algorithms for path finding, connectivity, ranking, and flow analysis. Matrices make it easier to apply algebraic methods to network structures.
7 Advanced topics
More advanced matrix topics extend the basic theory to specialized structures and broader operator concepts. These areas are important in higher mathematics and advanced applications.
7.1 Block matrices
A block matrix is divided into submatrices arranged in a larger rectangular pattern. The blocks may themselves be matrices of various sizes.
Block structure is useful when a problem naturally decomposes into parts. It often simplifies notation and allows calculations to be organized at a higher level of abstraction.
7.2 Sparse matrices
A sparse matrix contains many zero entries and relatively few nonzero ones. Such matrices arise frequently in large-scale scientific and engineering problems.
Sparse methods store and process only the nonzero entries, improving efficiency in memory use and computation. This is especially important for large systems with structured interactions.
7.3 Orthogonal and unitary matrices
An orthogonal matrix is a real square matrix whose transpose is also its inverse. A unitary matrix is the complex analogue, using the conjugate transpose.
These matrices preserve lengths and angles in their respective settings. They are important in numerical algorithms, geometry, quantum theory, and signal processing.
7.4 Positive definite matrices
A positive definite matrix is a symmetric matrix with a strictly positive quadratic form for every nonzero vector. Such matrices have strong stability and optimization properties.
They arise in covariance analysis, energy models, and constrained optimization. Positive definiteness also guarantees the existence of certain decompositions and uniqueness results.
7.5 Matrix functions
Matrix functions extend ordinary functions to matrices, allowing expressions such as powers, exponentials, and logarithms of matrices in appropriate contexts. These definitions are often built from series, diagonalization, or spectral methods.
Matrix functions are important in differential equations, system dynamics, and theoretical physics. They provide a way to apply analytic tools to linear operators and transformation matrices.