Analisis Metode Pengenalan Wajah Two Dimensial Principal Component Analysis (2DPCA) dan Kernel Fisher Discriminant Analysis Menggunakan Klasifikasi KNN (K- Nearest Neighbor)

Pengenalan wajah manusia (face recognition) merupakan salah satu bidang penelitian yang penting dan belakangan ini banyak aplikasi yang menerapkannya, baik di bidang komersil ataupun di bidang penegakan hukum. Pengenalan wajah merupakan sebuah sistem yang berfungsikan untuk mengidentifikasi berdasarkan ciri-ciri dari wajah seseorang berbasis biometrik yang memiliki keakuratan tinggi. Pengenalan wajah dapat diterapkan pada sistem keamanan. Banyak metode yang dapat digunakan dalam aplikasi pengenalan wajah untuk keamanan sistem, namun pada artikel ini akan membahas tentang dua metode yaitu Two Dimensial Principal Component Analysis dan Kernel Fisher Discriminant Analysis dengan metode klasifikasi menggunakan K-Nearest Neigbor. Kedua metode ini diuji menggunakan metode cross validation. Hasil dari penelitian terdahulu terbukti bahwa sistem pengenalan wajah metode Two Dimensial Principal Component Analysis dengan 5-folds cross validation menghasilkan akurasi sebesar 88,73%, sedangkan dengan 2-folds validation akurasi yang dihasilkan sebesar 89,25%. Dan pengujian metode Kernel Fisher Discriminant dengan 2-folds cross validation menghasilkan akurasi rata rata sebesar 83,10%.

Fracture identification in tight reservoirs by multiple kernel Fisher discriminant analysis using conventional logs

Kernel Fisher discriminant analysis (KFD) can map well log data into a nonlinear feature space to make a linear non-separable problem of fracture identification become a linear separable one. Commonly, KFD employs one kernel. However, the prediction capacity of KFD based on one kernel is limited to some extent, especially for a complex classification problem, such as fracture identification in tight sandstone reservoirs. To alleviate this problem, we employed a multiple kernel Fisher discriminant analysis (MKFD) method to recognize fracture zones. MKFD utilizes multi-scaled Gaussian kernel functions instead of a single kernel to realize the optimal nonlinear mapping. To assess the effectiveness of MKFD in fracture identification for complex reservoirs, we chose a dataset from tight sandstone reservoirs in China to implement comparison experiments. In the experiments, we used the MKFD with ten Gaussian kernels to map the original well logs into nonlinear feature spaces so that we could obtain appropriate features for fracture identification. The comparison results demonstrated that the accuracy of fracture identification by MKFD improved about 13.4% over KFD and MKFD also outperformed KFD in the blind well test, although the improvement of generalization ability of MKFD was not very obvious. Overall, MKFD can provide an accurate means for the identification of fracture zones in tight reservoirs. In this work, we also summarized the problems for fracture identification by MKFD.