| Summary: | The delphinidae family belongs to the Cetacean Order and is a member of the Odontocetes. The delphinidae family has characterized by the physical characteristics and frequency ranges of sound signals produced. Living in the sea and being a rare animal makes delphinidae very difficult to find and if we want to be classified, we have to capture and analyze the physicality of delphinidae. By using the fractal dimension we can analyze the sounds of the delphinidae family based on the characteristics of their sound signals to classify them. In this research, members of the delphinidae family will be classified using the Higuchi and K-Nearest Neighbor methods. By using 80 data, namely Common Dolphin 18 data, Killer Whale 20 data, Fraser's Dolphin 20 data, and Long-Finned Pilot Whale 22 data, the data used is .wav. In the first step, the Pre-Processing process will be carried out, then the feature extraction process will be carried out using the Discrete Wavelet Transform type mother wavelet Daubechies db4 wavelet with level 5 decomposition and Fast Fourier Transform. Then we will find the fractal dimension value using the Higuchi method. After obtaining the fractal dimension, the data will be divided into Training data and Testing data using k-cross validation with k value experiments namely 2, 4, 8, and 10. After the data is divided the data will be classified using K-Nearest Neighbor with an experimental K value. namely 1, 3, 5 and 7. In this study, the highest accuracy value was 82.5% with Kmax = 50, k = 8, and K = 3.Thus it can be concluded, the Higuchi and K-Nearest Neighbor methods can be used to classify members of the family delphinidae Keywords: Delphinidae, classification of member of family delphinidae, Higuchi fractal dimension, KNN
|
|---|