Multi-classification of Alcohols using Quartz Crystal Microbalance Sensors based-on Artificial Neural Network Single Layer Perceptron
Ferry Wahyu Wibowo, Wihayati Wihayati
Abstract
Detection of alcohol content has many benefits and plays a significant role in industry and product development research. Precise detection on a sensor can not classify at once, but the results of the sensor detection require further processing to determine the appropriate type of alcohol. One of the sensors that can detect types of alcohol is a quartz crystal microbalance (QCM). This paper uses an artificial neural network single-layer perceptron (ANN-SLP) model in processing datasets to determine multiple classifications of alcohol types. The ANN-SLP model uses a learning rate, Adam optimizer, and one-hot encoding. The types of alcohol used in the category in this paper are 1-Octanol, 1-Propanol, 2-Butanol, 2-Propanol, and 1-Isobutanol. The evaluation of the model uses different learning rates, namely 0.01 and 0.1 with 5000 epochs each. The best result for learning rate 0.01 is QCM3 sensor, while for learning rate 0.1 is QCM6. Both sensors have a test prediction accuracy of 100%.