Table 1 Advantages and disadvantages of ML algorithm
Content | |
|---|---|
Advantages | It can process a big data It can process a nonlinear data It can be used to select meaningful predictive variables and extract radiomic variables It usually has higher predictive performance than traditional statistical model |
Disadvantages | The accurate selection of ML algorithms is a challenge to establish a predictive model The design of ML predictive models lacks standards It is difficult to identify the process of ML model development due to the existence of “black box” The generalization ability of the established model needs to be further confirmed in validation cohort |