Figure 2.
t-SNE visualization of the training notes in the cohort 1 of eight I-codes based on TF-IDF
(A) t-SNE visualization of the TF-IDF word vectors of the cohort 1 training notes of eight I-code classification tasks.
(B) Prevalence of eight I-codes in the two selected regions with high prevalence of I-codes. Here, more than one I-code means all types of I-codes, not limited to the eight I-codes we investigated.