Klasifikasi Tingkat Kemurnian Bahan Bakar Minyak Berdasarkan Cepat Rambat Gelombang Menggunakan Algoritma K-Nearest Neighbor

Abstract

The need for fuel oil has increased along with the increase of population, the number of vehicles and industries. An increase in demand for fuel oil is used by some people to make a profit by selling mixed fuel oil at the same price as the price set by the government. The purpose of this study is to create a prototype device that can characterize the type of fuel oil and create a classification system to determine the level of fuel purity with 40 kHz ultrasonic waves based on the parameters of wave velocity using the K-Nearest Neighbor (KNN) algorithm.This device works by using a 40 kHz ultrasonic wave that is connected to an ultrasonic transmitter. The propagated wave will be received by the ultrasonic receiver. The wave received by the receiver will be amplified and connected to the comparator circuit so that it can be processed by a microcontroller. Data obtained using this tool are wave travel time, wave velocity, density, and attenuation. The data used for classification systems using the KNN algorithm is wave velocity.Classification using the KNN algorithm can identify the level of fuel purity based on the parameters of the wave velocity obtained from ultrasonic wave gauges with an accuracy of 72.50%. Wave velocity which is measured using ultrasonic waves is directly proportional to the actual speed with the largest percentage of deviations that is 0.34%.

Music Genre Identification Using SVM and MFCC Feature Extraction

“…Sinyal dinyatakan dalam skala MEL, skala ini didasarkan pada persepsi pitches dalam jarak interval yang sama yang dinilai oleh pengamat. Skala ini menggunakan filter yang ditempatkan secara linier pada frekuensi di bawah 1000 Hz dan jarak logarithmic di atas 1000Hz [6]. Bentuk diagram alir proses ekstraksi ciri MFCC ditunjukkan pada Gambar 2.…”