Research Paper
Year: 2020 | Month: April | Volume: 7 | Issue: 4 | Pages: 346-357
Genetic Algorithm Optimization Through Handling of Incomplete Data and Reduction
Sri Novida Sari1, Pahala Sirait2, Arwin Halim2
1Postgraduate Student at STMIK Mikroskil, Medan, Indonesia, 20212
2Lecturer Student at STMIK Mikroskil, Medan, Indonesia, 20212
Corresponding Author: Sri Novida Sari
ABSTRACT
Cervical cancer is a leading gynecologic malignancy worldwide. There are still many weaknesses in the data processing system in handling risk factors for cervical cancer. this research presents optimization techniques and shows the selection of features for the best combination of attributes in the risk of cervical cancer. There are thirty two attributes with eight hundred fifty eight samples. In addition, this data also has incomplete values due to respondents' privacy issues and imbalance data. Therefore, preprocessing techniques are used with regression imputation and data normalization methods. Furthermore, attribute reduction techniques with the optimum index factor (OIF) method are needed to improve the accuracy of the results. The results of research using the Genetic Algorithm method show that the comparison of optimization with reduction and without reduction of attributes has a difference of 1%. Optimization with no attribute reduction results in 95% while results with attribute reduction are 94%.
Keywords: Preprocessing, Incomplete data, regression imputation, optimum index factor, Genetic Algorithm, Optimization.
[PDF Full Text]