“…13 FTIR spectroscopy has been widely applied in various research studies including plants, 14,15 animals, 16–18 microorganisms, 19,20 food 21–23 and human diseases, 24–26 for various types of samples including serum, 27,28 urine, 29,30 cell lines, 31 and tissue sections. 32,33 Although a wide range of biological studies have been covered by FTIR spectroscopy, most of the analyses are based on the full spectrum, peak frequency shift 34 and intensity of the bands, 35,36 the relative ratios 37–39 of the vibrating groups and so on, 40 while in this study, we found that the similarity of the full spectrum between the HCC foci and the paracancerous tissue makes it difficult to identify the groups by some machine learning techniques, and hence, the peak areas of the spectra were calculated and selected as the features to perform the classification models of k -nearest neighbor (KNN), random forest (RF) and support vector machine (SVM). These models were evaluated through 5-fold cross-validation, and the index of precision, recall and F 1-score were used to characterize the performance.…”