Fig 1.
Preprocessing flow chart of “quanteda” natural language processing package.
Fig 2.
Difference of the text character lengths between AIS and non-AIS reports.
AIS, acute ischemic stroke.
Fig 3.
Result of keyness plot analysis of AIS and non-AIS reports.
AIS, acute ischemic stroke.
Fig 4.
Comparison of ML and NLP algorithms for classifying the brain MRI reports.
ML, machine learning; NLP, natural language processing, BLR, binary logistic regression; NBC, naïve Bayesian classification; SDT, single decision tree; SVM, support vector machine; TFIDF, term frequency-inverse document frequency.
Table 1.
Results of performance of each machine learning algorithms.
Table 2.
Error analysis of result of single decision tree in classifying AIS and non-AIS.