@article{IPI1261999, title = "KLASIFIKASI AUDIO MENGGUNAKAN WAVELET TRANSFORM DAN NEURAL NETWORK", journal = "Universitas Janabadra", volume = "Vol 4, No 2 (2019): Jurnal Informasi Interaktif", pages = "", year = "2019", url = http://e-journal.janabadra.ac.id/index.php/informasiinteraktif/article/view/875/596 author = "Yulianto Mustaqim; Ema Utami; Suwanto Raharjo", abstract = "Biodiversity that exists in nature shows the overall variation between living things both from the smallest levels, namely genes, species and eskosistem. One animal with a fairly high level of variation, namely birds chirping. Chirping has an identifier for each type both of the color of the feather, body shape, shape of the beak, food, how to find food and the most obvious is the difference in the chirping of birds. The problem faced is the number of species of birds chirping that are almost similar to each other so the introduction of birds with sound becomes quite difficult. This makes the introduction of birds with sound requires a special technique. The techniques used are transform wavelets and neural networks. At the end of the study, obtained Wavelet Package Decomposition extraction with training data used as many as 500 data. There are two preprocessing methods that are done by cutting and resampling (downsampling). The most optimal number of neurons to be used in hidden layers is 256 neurons with 500 epochs. The highest accuracy is 88.6% with momentum 0.2, learning rate 0.2 and wavelet daubechies2 while the lowest accuracy is 74.2% with momentum 0.8, learning rate 0.8 and wavelets haar.  Keywords: Classification, Neural Network, Wavelet Transform, Haar, Daubechies2", }