Intelligent Medical Disease Diagnosis Using Improved Hybrid Genetic Algorithm - Multilayer Perceptron Network

摘要: An improved genetic algorithm procedure is introduced in this work based on the theory of most highly fit parents (both male and female) are likely to produce healthiest offspring. It avoids destruction near optimal information promotes further search around potential region by encouraging exchange important among fittest solution. A novel crossover technique called Segmented Multi-chromosome Crossover also introduced. maintains contained gene segments allows offspring inherit from multiple parent chromosomes. The GA applied for automatic simultaneous parameter optimization feature selection multi-layer perceptron network medical disease diagnosis. Compared previous works, average accuracy proposed best all algorithms diabetes heart dataset, second cancer dataset.