Research Article
A Robust Hybrid Approach Based on Estimation of Distribution Algorithm and Support Vector Machine for Hunting Candidate Disease Genes
Algorithm 1
The step-by-step recipe for the computational algorithm of the EDA-SVM approach.
Step 1. Read gene expression profile matrix from database, is the number of genes in . | Step 2. Generate individuals (the initial population) randomly. Each individual has an - | length vector of bits of either 1 or 0. | Step 3. For each individual in , determine: | a gene subset corresponding to individual . If bit i equals to 1, include in the subset. | gene expression profile submatrix. | . | Step 4. retain individuals with the highest evaluations. | Step 5. arg calculate marginal distribution of variable of bit i based on by | using the formula: , where is the value of the variable in | individual . |
calculate weight of corresponding to feature based on . | , where is weight of bit in | individual . | compute probability distribution of each bit , which is written | mathematically as: | PrPr. | is learning rate. is generated at random. | Step 6. generate new individuals by sampling the probability distribution. | Step 7. . | Step 8. . | Step 9. Endoutput the optimal individual based on the evaluation with: . |
|