Co-occurrence Matrices and PMI-SVD Embeddings (CRAN, 2025)

The nlpembeds package provides efficient methods to compute co-occurrence matrices, pointwise mutual information (PMI) and singular value decomposition (SVD) embeddings. In the biomedical and clinical setting, one challenge is the huge size of databases of electronic health records (EHRs), e.g. when analyzing data of millions of patients over tens of years. To address this, this package provides functions to efficiently compute monthly co-occurrence matrices, which is the computational bottleneck of the analysis, by using the RcppAlgos package and sparse matrices. Furthermore, the functions can be called on SQL databases, enabling the computation of co-occurrence matrices of tens of gigabytes of data, representing millions of patients over tens of years. PMI-SVD embeddings are extensively used, e.g. in Hong C. (2021).