Paper Title
Applying Genetic Network Programming For A Nude Image Classification

Abstract
According to Parental Control data analysis from Kaspersky Lab in 2003, Pornographic and erotic websites (16.83%) are most attractive for children in the Internet. More than half (59.5%) of users encountered pornography in 2015 in the same study. Moreover, most of nude images such as celebrity are made without a personís consent. There are two classes including nude or not nude images; therefore, the traditional decision tree (ID3) technique is suited for this research. The traditional decision tree such as ID3 theory is used to construct a simple decision tree; however, nude image classification is NP hard problems. A genetic algorithm is applied to help to improve an accuracy of classification. In this paper, the performance of the proposed algorithm is evaluated by comparing it with the traditional decision tree classification and genetic network programming (GNP) algorithm. With a traditional decision tree, the accuracy result is 85.2 %. The reason why the accuracy result is not so high is because some rules is pruned with ID3 algorithm. Genetic algorithm is good at exploring all possible answers in the universe. The scope of searching depends on a fitness function. Mutation and cross over can increase some improvement. The proposed method with mutation and crossover is up to 95.61%. Keywords- Data Mining, Genetic Network Programming, Nude Image, Image Processing