A repository for final project for POLS 904 at University of Kansas, FS 17, Paul Johnson professor.

This project demonstrates the feasability of using sparse matrices to improve management and performance in R. Specifically, I modify and extend the capabilities of the {pcse} package by implementing sparse capabilities provided by the {Matrix} package. I identify a recurring issue in processing large N x T matrices in which the package fails to produced the required Kronecker multiplication and crashes R. Then, I identify the matrix storage structure used in the original package, which I hypothesize causes memory overload. I provide background on the sparse matrxi alternative, and provide examples of implementation. I find notable improvements in both dimensional size and speed capabilities.