Table 2.

Literature review

Study Name and Year	Medical Vertical	Data	ML Method	Result
Fujisawa et al. (2019)[6]	Skin cancer classification	6,009 clinical images	Deep Convolutional Neural Network	Accuracy: 76.5% Sensitivity: 96.3% Specificity: 89.5%
Esteva et al. (2017)[3]	Classification of skin lesions	129,450 clinical images	Convolutional Neural Network	AUC between 0.91 and 0.96
Masood et al. (2013)[7]	Skin cancer	Skin images	Support Vector Machine, Discriminant Analysis, Decision Trees, Logistic Regression	Varying levels of accuracy, often not benchmarked against one another or a reference data set. Need for better validation data was pointed out. Note: studies included from as early as 1993
Fjell et al. (2009)[8]	Identifying, new antibacterial agents	silico library of approximately 100,000 peptides	Artificial Neural Networks	Accuracy: 94%