Multilayer Restricted Boltzmann Machine Research Articles

Improved RBM‐based feature extraction for credit risk assessment with high dimensionality

AbstractTo address the high‐dimensional issues in credit risk assessment, an improved multilayer restricted Boltzmann machine (RBM) based feature extraction method is proposed. In the improved multilayer RBM methodology, the reconstruction error method is first applied to ensure the number of RBM layers to construct an optimal model and then the weighted pruning approach is used to remove redundant and irrelevant traits. For verification purposes, two real‐world credit datasets are employed to demonstrate the effectiveness of the proposed multilayer RBM methodology. The experimental results reveal that a significant improvement in credit classification performance can be obtained by the improved multilayer RBM methodology. This indicates the improved multilayer RBM model proposed in this paper can be used as a promising tool to solve the high‐dimensionality issues in credit risk evaluation.

Compilation of cutting load spectrum for CNC lathe based on DBN-BP prediction model

Load spectrum can effectively represent variations in statistical load, and it is the foundation data for the overall product’s reliability test, reliability design and fatigue life analysis. However, compiling a load spectrum of small sample conditions is difficult because of the complicated and variable load conditions of machine tools and kaleidoscopic cutting process. Additionally, extreme load extrapolating is susceptible to subjective influences, thereby resulting in some uncertainties. Therefore, this study proposes a cutting load spectrum compilation method for computerised numerical control (CNC) lathe based on deep belief network-back propagation (DBN-BP) load prediction model, combining multi-source load condition information with deep learning. Firstly, single-layer BP and multi-layer restricted Boltzmann machine (RBM) were used to create a cutting load prediction model. Mixed distribution of mean and range of dynamic cutting load and parameters in copula function are predicted by using cutting process parameters. The number of hidden layers and single-layer nodes in the prediction model were determined using mean average (MA) precision and training time. Secondly, dynamic cutting load conditions were divided into the prediction model’s training, test and prediction sets; and the model’s accuracy was verified using comparative tests. Lastly, the two-dimensional mean-range matrix of load in each typical working condition was compiled using the prediction model, and the cutting load spectrum and two-dimensional CNC lathe programme loading spectrum were constructed. The proposed method saves time and personnel for numerous cutting tests when working with small samples and also assures load prediction accuracy.

Dropout Deep Belief Network Based Chinese Ancient Ceramic Non-Destructive Identification.

A non-destructive identification method was developed here based on dropout deep belief network in multi-spectral data of ancient ceramic. A fractional differential algorithm was proposed to enhance the spectral details by making use of the difference between the first and second-order differential pre-process spectral data. An unsupervised multi-layer restricted Boltzmann machine (RBM) was employed to extract some high-level features during pre-training. Some weight and bias values trained by RBM were used to initialize a back propagation (BP) neural network. The RBM deep belief network was fine-tuned by the BP neural network to promote the initiative performance of network training, which helped to overcome local optimal limitation of the network due to the random initializing weight parameter. The dropout strategy has been put forward into the RBM network to solve the over-fitting of small sample spectral data. The experimental results show that the proposed method has excellent recognition performance of the ceramics by comparisons with some other ones.

The promotion strategy of supply chain flexibility based on deep belief network

Supply chain flexibility is the processing ability of the enterprize to deal with the uncertain environment of supply and demand. In this paper, we consider the supply side (node interrupt) and demand side (demand fluctuations) under uncertain environment. By using the deep belief network (DBN), which is composed of multilayer Restricted Boltzmann Machine (RBM), it establishes the supply chain flexibility network with optimization of the transfer node and flow. The deep belief network is trained by the data of large manufacturing enterprize, compared with the traditional neural network (MLR, BP and GA). The results show that the deep belief network model overcomes the shortcomings of the traditional neural networks, such as easy to fall into local optimum, long training time and low function fitting degree, and it has higher prediction accuracy. This network model based on the deep belief network can promote the supply chain flexibility more, when supply and demand fluctuations occur.

A new item-based deep network structure using a restricted Boltzmann machine for collaborative filtering

The collaborative filtering (CF) technique has been widely used recently in recommendation systems. It needs historical data to give predictions. However, the data sparsity problem still exists. We propose a new item-based restricted Boltzmann machine (RBM) approach for CF and use the deep multilayer RBM network structure, which alleviates the data sparsity problem and has excellent ability to extract features. Each item is treated as a single RBM, and different items share the same weights and biases. The parameters are learned layer by layer in the deep network. The batch gradient descent algorithm with minibatch is used to increase the convergence speed. The new feature vector discovered by the multilayer RBM network structure is very effective in predicting a rating and achieves a better result. Experimental results on the data set of MovieLens show that the item-based multilayer RBM approach achieves the best performance, with a mean absolute error of 0.6424 and a root-mean-square error of 0.7843.

A Multilayer Improved RBM Network Based Image Compression Method in Wireless Sensor Networks

The processing capacity and power of nodes in a Wireless Sensor Network (WSN) are limited. And most image compression algorithms in WSN are subject to random image content changes or have low image qualities after the images are decoded. Therefore, an image compression method based on multilayer Restricted Boltzmann Machine (RBM) network is proposed in this paper. The alternative iteration algorithm is also applied in RBM to optimize the training process. The proposed image compression method is compared with a region of interest (ROI) compression method in simulations. Under the same compression ratio, the qualities of reconstructed images are better than that of ROI. When the number of hidden units in top RBM layer is 8, the peak signal-to-noise ratio (PSNR) of the multilayer RBM network compression method is 74.2141, and it is much higher than that of ROI which is 60.2093. The multilayer RBM based image compression method has better compression performance and can effectively reduce the energy consumption during image transmission in WSN.