Comment on Finding reduced Raman spectroscopy fingerprint of skin samples for melanoma diagnosis through machine learning

This paper comments on the article Finding reduced Raman spectroscopy fingerprint of skin samples for melanoma diagnosis through machine learning by D.C. Araujo et al. The authors apply Raman spectroscopy for the classification of benign and malignant skin neoplasms based on their Raman spectra. Despite the high performance of the proposed technique it may provide unreasonably high accuracy because of incorrect cross-validation procedure. To confirm the possibility to discriminate neoplasm skin tissues based on Raman spectra analysis the authors should provide additional data regarding utilized cross-validation procedure.

Automated summary

This paper comments on a study that uses machine learning to find reduced Raman spectroscopy fingerprint of skin samples for melanoma diagnosis. The authors classify benign and malignant skin neoplasms based on their Raman spectra. However, the proposed technique may have unreasonably high accuracy due to incorrect cross-validation procedure. Additional data regarding the utilized cross-validation procedure should be provided to confirm the possibility of discriminating neoplasm skin tissues based on Raman spectra analysis.