Business Mathematics (BK/IBA) Quantitative Research Methods I (EBE) Matrices BUSM/QRM1 1 Matrices Contents • Matrices • Special matrices • Operations with matrices • Matrix multipication • Matrix transposition • Symmetric matrices BUSM/QRM1 2 Matrices Matrices • A matrix is a rectangular array of numbers or variables • Notation • we often use roman bold capital letters to refer to them , , ⋯ , 3 2 , , ⋯ , • e.g., = −2 ,= 0 ⋯ ⋯ ⋯ ⋯ 12.5 −12.7 , , ⋯ , • Terminology • • • • BUSM/QRM1 these matrices consist of 6 respectively elements the order (or size is) 3 × 2 respectively × when = , the matrix is a square matrix when ≠ , the matrix is rectangular 3 Matrices Matrices • Notice the order of the indices 3 2 • In the matrix = −2 the element , refers 0 12.5 −12.7 to the cell at row 2 and column 1 • so to −2 • while , is in row 1 and column 2 and has value 2 • The order of is 3 × 2, not 2 × 3 • Convention : • !,"#$% ! , × "#$% • when no ambiguity you may skip comma: &' instead of &,' BUSM/QRM1 4 Matrices Special matrices • Zero matrix: ( = 0 0 ⋯ 0 0 0 ⋯ 0 ⋯ 0 ⋯ 0 ⋯ ⋯ ⋯ 0 • for a matrix of any order • Identity matrix: ) = • for a square matrix BUSM/QRM1 1 0 ⋯ 0 0 1 ⋯ 0 ⋯ ⋯ ⋯ ⋯ 0 0 ⋯ 1 5 Matrices Operations with matrices • We can define some basic operations with matrices, similar to the basic operations with vectors • • • • • addition ( + +, through + , &' = &' + ,&' ) multiplication (-, through - &' = - × &' ) negative matrix (−, through − &' = −&' ) subtraction ( − +, through − , &' = &' − ,&' ) equality ( = +, through &' = ,&' ) • But what about the inner product? • not available for matrices • instead: matrix multiplication BUSM/QRM1 6 Matrices Matrix multiplication • Let and + be two matrices, of order × . respectively .× • We define the matrix product + as + 1 &' = / &0 ,0' , 3 = 1, … , , 5 = 1, … , 02 • alternatively written as ⋅ + • but do not write × + • Notice well: the result of the multiplication of two matrices is a matrix • different for the inner product of two vectors BUSM/QRM1 7 Matrices Matrix multiplication • Illustration: • + • + BUSM/QRM1 ∑ = , 02 ,0 ,0, = , ,, + , ,, 8,8 = ∑02 8,0 ,0,8 = 8, ,,8 + 8, ,,8 8 Matrices Matrix multiplication • Notice well the orders of the matrices: • 1 + 1 = + • so #columns in should match #rows in + • and #rows in + is #rows in • and #columns in + is #columns in + • Consequences: given a matrix or order 3 × 3 and a matrix + of order 3 × 2 • + exists and is of order 3 × 2 • + does not exist • what about ? and ++? and + + ? BUSM/QRM1 9 Matrices Matrix multiplication • It follows that (with suitable , +, and 9) • + + 9 = + + 9 (distributive property) • + 9 = +9 = +9 (associative property) • But not that • + = + (commutative property) 1 2 2 −8 and + = • example: take = 3 6 −1 4 • + = 0 0 −22 −44 , but + = 11 22 0 0 • By the way, notice that: • + = (, while ≠ ( and + ≠ ( • while for numbers , = 0 ⇔ = 0or, = 0 BUSM/QRM1 10 Matrices Matrix multiplication • Some properties (for suitable , +, and 9): • ( = ( and ( = ( • ) = and ) = • + = 9 ⇏ + = 9 • example: = • + = 9 = BUSM/QRM1 1 2 3 −4 1 4 ,+= , and 9 = 3 6 −2 3 −1 −1 −1 2 −3 6 11 Matrices Matrix multiplication • What about powers of a matrix? • Let us define = for any square matrix • why square? • mind the differenve between “a square matrix” and “a squared matrix” • Likewise 8 = , etc. • what is : is it or is it ? • More in BUSM/QRM1 general = @ A 12 =1 = 2,3, … Matrices Matrix transposition • Consider = × = , , ⋯ , , , ⋯ , ⋯ , ⋯ , ⋯ ⋯ ⋯ , • The transpose of , denoted by B is given by , , ⋯ , , , ⋯ , B = ⋯ ⋯ ⋯ ⋯ , , ⋯ , • In words, B has rows and Bcolumns so B is a ( × )-matrix and row 3 of is column 3 of • “reflection in the diagonal” BUSM/QRM1 13 Matrices Matrix transposition • Some properties (for suitable , +, and 9): • B B = • + + B = B + + B • + B = + B B and +9 B = 9B + B B • - B = -B BUSM/QRM1 14 Matrices Symmetric matrices • Definition • The matrix is symmetric if and only if = B • so if and only if &' = '& for all 3, 5 • note: only a square matrix can be symmetric 1 3 • Example: = is symmetric 3 6 1 2 −1 • If = the B is symmetric 3B 6 5 B • note: is symmetric too but in general ≠ B (you can see this for the example without even doing a calculation!) • In general B is symmetric for an arbitrary matrix • and so is B (why?) BUSM/QRM1 15 Matrices