斯坦福大学机器学习课程讲义第三讲―― 线性代数基础

AndrewNgLinearAlgebrareview(optional)MatricesandvectorsMachineLearningAndrewNgDimensionofmatrix:numberofrowsxnumberofcolumnsMatrix:Rectangulararrayofnumbers:AndrewNgMatrixElements(entriesofmatrix)“,entry”intherow,column.AndrewNgVector:Annx1matrix.n-dimensionalvector1-indexedvs0-indexed:elementAndrewNgAndrewNgLinearAlgebrareview(optional)AdditionandscalarmultiplicationMachineLearningAndrewNgMatrixAdditionAndrewNgScalarMultiplicationAndrewNgCombinationofOperandsAndrewNgAndrewNgLinearAlgebrareview(optional)Matrix-vectormultiplicationMachineLearningAndrewNgExampleAndrewNgDetails:mxnmatrix(mrows,ncolumns)nx1matrix(n-dimensionalvector)m-dimensionalvectorToget,multiply’srowwithelementsofvector,andaddthemup.AndrewNgExampleAndrewNgHousesizes:AndrewNgAndrewNgLinearAlgebrareview(optional)Matrix-matrixmultiplicationMachineLearningAndrewNgExampleAndrewNgDetails:mxnmatrix(mrows,ncolumns)nxomatrix(nrows,ocolumns)mxomatrixThecolumnofthematrixisobtainedbymultiplyingwiththecolumnof.(for=1,2,…,o)AndrewNgExampleAndrewNgHousesizes:MatrixMatrixHave3competinghypotheses:1.2.3.AndrewNgAndrewNgLinearAlgebrareview(optional)MatrixmultiplicationpropertiesMachineLearningAndrewNgLetandbematrices.Theningeneral,(notcommutative.)E.g.AndrewNgLetLetComputeComputeAndrewNgIdentityMatrixForanymatrix,Denoted(or).Examplesofidentitymatrices:2x23x34x4AndrewNgAndrewNgLinearAlgebrareview(optional)InverseandtransposeMachineLearningAndrewNgNotallnumbershaveaninverse.Matrixinverse:IfAisanmxmmatrix,andifithasaninverse,Matricesthatdon’thaveaninverseare“singular”or“degenerate”AndrewNgMatrixTransposeExample:Letbeanmxnmatrix,andletThenisannxmmatrix,and