Articles

Found 2 Documents
Search
Journal : INTI Nusa Mandiri

PENERAPAN KLASIFIKASI ALGORITMA C4.5 PADA FITUR GRAY LEVEL CO-OCCURRANCE MATRIX UNTUK ANALISA TEKSTUR CITRA WAJAH Kurniawan, Ilham; Riana, Dwiza
INTI Nusa Mandiri Vol 14 No 1 (2019): INTI Periode Agustus 2019
Publisher : Pusat Penelitian dan Pengabdian Pada Masyarakat

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (1023.937 KB)

Abstract

Research on facial images is useful to distinguish the characteristics of each human being. The introduction of healthy and unhealthy facial skin images aims to identify human skin types automatically. For this purpose features such as contrast, correlation, energy, homogeneity, which are features of the Gray Level Co-occurrance Matrices (GLCM) are used. This study proposes a method for analyzing and classifying the GLCM texture on facial skin. The image used in this study was taken in the face image section which consists of the skin of the cheek and the whole face. The methods used are image acquisition, facial skin image, ROI selection, RGB image conversion to gray image, GLCM feature extraction, C4.5 algorithm classification and evaluation. The results showed that the C4.5 algorithm classification on texture analysis of facial images produced an accuracy value of 66.67%, the accuracy value was still low and the need for further research could not be used to increase the accuracy of texture analysis of facial images.
INTEGRASI METODE SAMPLE BOOTSTRAPPING DAN WEIGHTED PRINCIPAL COMPONENT ANALISYS (PCA) UNTUK MENINGKATKAN PERFORMA NAÏVE BAYES PADA CITRA TUNGGAL PAP SMEAR Dewi, Yumi Novita; Rianto, Harsih; Riana, Dwiza; Siregar, Juarni
INTI Nusa Mandiri Vol 14 No 2 (2020): INTI Periode Februari 2020
Publisher : Pusat Penelitian dan Pengabdian Pada Masyarakat

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.33480/inti.v14i2.1103

Abstract

Research on cervical cancer with the Pap Smear method is useful for finding pre-cancer diagnoses. Associated with previous research that the accuracy of the Naïve Bayes algorithm to the classification of a single Pap smear image still has an unsatisfactory accuracy. Whereas determining the class of single Pap cell smears is very important in determining whether these cells are normal or not. This study aims to determine whether integration using the Sample Bootstrapping (SB) method with the Weighted Principal Component Analysis (W-PCA) algorithm can improve the performance of the Naïve Bayes algorithm for seven different cell types. This model is the best solution used in the classification of datasets that are classified as having large dimensions. So that the integration of the two algorithms can increase the accuracy value to 87.24% for the seven classes and 97.30% for the two classes, and it can be concluded that with this integration model can improve the best accuracy value.