IDENTIFIKASI DAGING AYAM LAYAK KONSUMSI DENGAN HIDUNG ELEKTRONIK MENGGUNAKAN JARINGAN SYARAF TIRUAN MODEL MULTI LAYER PERCEPTRON; IDENTIFICATION OF DECENT CONSUMPTION CHICKEN USING ELECTRONIC NOSE WITH NEURAL NETWORK MULTI LAYER PERCEPTRON MODEL

SRI SUWARYANA PUTRA, Danang Lelono

2015 | Skripsi | FMIPA

Broiler chicken is a food that contains high nutrition. Chicken has weakness, that is easily decomposed. In the decay of chicken protein decomposition by bacteria that produce indole compounds, skatol, H2S and mercaptans aminamin and NH3. This scent can be used for detection of eligibility consumption of chicken. Chicken samples were grouped with variations of chicken into 1, 2, 4, and 8 days. Chicken fit for consumption if aged no more than 2 days, with sample mass is 25 grams. Data were collected using an electronic nose for 8 days by the sensing process and flushing each for 180 seconds. In this study, Processing data using Principal Component Analysis (PCA) and Neural Network Multi Layer Perceptron (ANN-MLP) with the aim to identify chicken based on time storage). The results of cumulative percentage variance of two main components to the testing classification between chicken unfit for consumption with chicken unfit for consumption is 97.6%, while the testing of test data using ANN MLP obtained an accuracy of 93.33%. Thus, we can conclude electronic nose using ANN-MLP and PCA can distinguish between chicken unfit for consumption with chicken unfit for consumption.

Kata Kunci : Chicken; electronic nose; artificial neural network; multi layer perceptron principal component analysis