Comparing Fitness Functions for Genetic Feature Transformation

Partial Functions in Fitness-Shared Genetic Programming

This paper investigates the use of partial functions and fitness sharing in genetic programming. Fitness sharing is applied to populations of either partial or total functions and the results compared. Applications to two classes of problem are investigated: learning multiplexer definitions, and learning (recursive) list membership functions. In both cases, fitness sharing approaches outperform...

Feature-Denoising Based on Average Fitness of Genetic Population

To solve the problem of the noise existed in feature items of category template in filtering system, weight adjusting strategy based on average fitness of population is proposed combining genetic algorithm with feedback. The feature items’ contribution to individual fitness is studied to adjust feature items’ weight by the genetic difference in the average fitness of the individual. Experimenta...

Feature Transformation: A Genetic-Based Feature Construction Method for Data Summarization

The importance of input representation has been recognized already in machine learning. This article discusses the application of genetic-based feature construction methods to generate input data for the data summarization method called Dynamic Aggregation of Relational Attributes (DARA). Here, feature construction methods are applied to improve the descriptive accuracy of the DARA algorithm. T...

Improving Genetic Algorithms' Efficiency Using Intelligent Fitness Functions

Genetic Algorithms are an effective way to solve optimisation problems. If the fitness test takes a long time to perform then the Genetic Algorithm may take a long time to execute. Using conventional fitness functions Approximately a third of the time may be spent testing individuals that have already been tested. Intelligent Fitness Functions can be applied to improve the efficiency of the Gen...

Efficient evaluation genetic algorithm under integrated fitness functions