Design Of Multilayer Perceptron Research Articles

Geometrical Interpretation and Design of Multilayer Perceptrons.

The multilayer perceptron (MLP) neural network is interpreted from the geometrical viewpoint in this work, that is, an MLP partition an input feature space into multiple nonoverlapping subspaces using a set of hyperplanes, where the great majority of samples in a subspace belongs to one object class. Based on this high-level idea, we propose a three-layer feedforward MLP (FF-MLP) architecture for its implementation. In the first layer, the input feature space is split into multiple subspaces by a set of partitioning hyperplanes and rectified linear unit (ReLU) activation, which is implemented by the classical two-class linear discriminant analysis (LDA). In the second layer, each neuron activates one of the subspaces formed by the partitioning hyperplanes with specially designed weights. In the third layer, all subspaces of the same class are connected to an output node that represents the object class. The proposed design determines all MLP parameters in a feedforward one-pass fashion analytically without backpropagation. Experiments are conducted to compare the performance of the traditional backpropagation-based MLP (BP-MLP) and the new FF-MLP. It is observed that the FF-MLP outperforms the BP-MLP in terms of design time, training time, and classification performance in several benchmarking datasets. Our source code is available at https://colab.research.google.com/drive/1Gz0L8AnT4ijrUchrhEXXsnaacrFdenn?usp = sharing.

Enhancement of ultrasonic-assisted extraction and antioxidant potential of phenolic compounds from Moroccan Cannabis sativa: A green intelligent approach integrating optimal mixture process design and artificial neural networks

Environmentally-friendly extraction methods enables both efficient extraction of phenolic compounds from plant sources and safe use of the extracted substances. These characteristics render them as ideal choices for sustainable extraction processes. This study combines a green and sustainable extraction strategy with an intelligent statistical approach to optimize the recovery and the antioxidant activity of phytochemicals derived from Moroccan Cannabis sativa. The optimization of the ultrasound-assisted extraction is achieved using an optimal mixture-process design (OMPD) combined with artificial neural networks (ANNs). The OMPD consists of a mixture of water, glycerol, and ethanol, along with process parameters such as processing time, solvent-to-material (S/M) ratio, and temperature. The objective is to optimize the five responses extraction yield, total phenolic content (TPC), and antioxidant activities measured by 2,2-diphenyl-1-picrylhydrazyl (DPPH), total antioxidant capacity (TAC), and Ferric reducing antioxidant power (FRAP) assays. The results of the OMPD indicate that the recovery of phenolic compounds and antioxidant activity can be improved by using a binary mixture of water and ethanol, along with moderate processing time and temperature, and a high S/M ratio. The optimal responses were 26.06 %, 92.6 mg GAE/g, 0.1 mg/ml, 298 mg AAE/g, and 1.3 mg/ml for yield, TPC, TFC, DPPHIC50, TAC, and FRAPEC50, respectively. Subsequently, the OMPD matrix was used as input for the ANNs, employing a 6-6-5 multi-layer perceptron design. The findings of the ANNs closely matched those obtained from the OMPD models. However, the performance parameters of the ANNs models were slightly superior to those of the OMPD.

Evaluation of an optimal design method for a multilayer perceptron by using the design of experiments

We evaluated the performance of an optimal design method for a multilayer perceptron (MLP) by using the design of experiments (DOE). In our previous work, we proposed an optimal design method for MLPs in order to determine the optimal values of such parameters as the number of neurons in the hidden layers and the learning rates. In this article, we evaluate the performance of the proposed design method through a comparison with a genetic algorithm (GA)-based design method. We target an optimal design of MLPs with six layers. We also evaluate the proposed designed method in terms of calculating the amount of optimization. Through the above-mentioned evaluation and analysis, we aim at improving the proposed design method in order to obtain an optimal MLP with less effort.