PENERAPAN METODE K-NEAREST NEIGHBOUR (K-NN) DALAM MENGELOMPOKKAN JENIS KALENG BERDASARKAN CITRA RGB
Daftar Isi:
- Principle of packaging in the modern world is upgrading by using the better technology, the type of packaging has variated from paper, plastic, glass, metal (tin), fiber and laminated materials. The most materials of metal that we used for packaging are tin-plate, aluminum, and aerosol. Metal are the recyclable materials, for the process of recycling firstly we cluster the metal by the type. In this study the types of metal will be clustering using K-Nearest Neighbor (K-NN) by using red, green, and blue (RGB) pixel data from metal using help from MATLAB software. External factors will be added to the RGB image as the testing data fot clustering, the process of selecting testing data that have most significant factors will be done by Random Block Design (RBD) method and reduced of independent variables of the data will use the Principle Component Analysis (PCA) using from Minitab software. The result of clustering the type of metal on this study indicate the most high accuracy of the clustering using the original training data without PCA and 16 original data testing without RBD and also clustering by using the training data the results of KU score and testing data of RBD result of KU score are 45.6% on testing data with factors type of lamp 1 angel of lighting 900 and speed of board 2.