PEMBUATAN SISTEM MACHINE LEARNING DENGAN METODE SUPPORT VECTOR MACHINE (SVM) YANG TER-COUPLED OLEH ELECTRONIC NOSE (E-Nose) UNTUK MENDETEKSI POLA AROMA KULIT BABI DAN KULIT SAPI

Desrinda Mala Dwi Putri, NIM.: 16620007 (2020) PEMBUATAN SISTEM MACHINE LEARNING DENGAN METODE SUPPORT VECTOR MACHINE (SVM) YANG TER-COUPLED OLEH ELECTRONIC NOSE (E-Nose) UNTUK MENDETEKSI POLA AROMA KULIT BABI DAN KULIT SAPI. Skripsi thesis, UIN SUNAN KALIJAGA YOGYAKARTA.

Preview	Text (PEMBUATAN SISTEM MACHINE LEARNING DENGAN METODE SUPPORT VECTOR MACHINE (SVM) YANG TER-COUPLED OLEH ELECTRONIC NOSE (E-Nose) UNTUK MENDETEKSI POLA AROMA KULIT BABI DAN KULIT SAPI) 16620007_BAB-I_IV-atau-V_DAFTAR-PUSTAKA.pdf Download (4MB) | Preview
	Text (PEMBUATAN SISTEM MACHINE LEARNING DENGAN METODE SUPPORT VECTOR MACHINE (SVM) YANG TER-COUPLED OLEH ELECTRONIC NOSE (E-Nose) UNTUK MENDETEKSI POLA AROMA KULIT BABI DAN KULIT SAPI) 16620007_BAB-II_sampai_SEBELUM-BAB-TERAKHIR.pdf Restricted to Registered users only Download (7MB)

Abstract

This study aimed to detect the aroma patterns of pigskin and cow hide using E-Nose and classify the aroma patterns using the ML-SVM coupled by E-Nose. This research was conducted in two stages, namely data collection and data processing. Data were collected using Electronic Nose (E-Nose) and GeNose datalogger software. Data processing was performed by using machine learning (ML) based on the support vector machine (SVM) algorithm. The results of the E-Nose detection dataset were pre-processed using feature extraction with the method of reducing the baseline value and the average value. The experimental dataset was divided into three subsets, namely training data, testing data and new-testing data. The classification of the aroma pattern of pigskin and cow-hide used ML with a non-linear supervised learning analysis method, SVM. In addition, hyperparameter optimization and K-fold cross-validation variants were implemented during the model training procedure to select the best classification model and to avoid over-fitting issues. The aroma patterns of pigskin and cow-hide have been detected and classified. The classification results of the SVM model obtained an accuracy of 100% for testing data and 100% for new-testing data.

Share this knowledge with your friends :

Actions (login required)

View Item