First Study of 169 Linearly Separable Expression Data

I studied ECG diagnosis using LDF after 1971. It was inferior to the doctor’s Decision Tree. This failure motivated my 53-year LSD study in Table 1. LDF could not find data-specific misclassified cases to confirm if the external criteria decided by doctors are correct. Therefore, I developed the combinatorial LDF named RIP to Minimizes Number of Misclassifications from all combination of n-cases and p-vars and completed the LSD theory using five regular and 169 gene expression datasets.