MIT OpenCourseWare
46 minutes
Dimension and Basis for Matrix Spaces from Symmetric Matrices to Small World Graphs
Linear algebra generalizes vector space concepts to abstract structures comprising elements that support addition and scalar multiplication but lack a specific geometric representation, such as matrices, functions satisfying linear differential equations, or weighted adjacency graphs. The core principle is the decomposition of these spaces into fundamental subspaces—specifically row, column, null, and left-null spaces—and their dimensions are strictly determined by rank formulas like $n - r = \dim(\text{Null})$ and $\dim(U) + \dim(V) = \dim(U+V) + \dim(U \cap V)$ for summing subspaces. This framework enables the analysis of structural properties, such as describing "small world" phenomena in graph theory, where the distance between two nodes is the fewest edges on a path between them, bridging pure algebraic manipulation with network science applications.