In this brief, the optimally pruned extreme learning machine (OP-ELM) methodology is presented. It is based on the original extreme learning machine (ELM) algorithm with additional steps to make it more robust and generic. The whole methodology is presented in detail and then applied to several regression and classification problems. Results for both computational time and accuracy (mean square error) are compared to the original ELM and to three other widely used methodologies: multilayer perceptron (MLP), support vector machine (SVM), and Gaussian process (GP). As the experiments for both regression and classification illustrate, the proposed OP-ELM methodology performs several orders of magnitude faster than the other algorithms used in this brief, except the original ELM. Despite the simplicity and fast performance, the OP-ELM is still able to maintain an accuracy that is comparable to the performance of the SVM. A toolbox for the OP-ELM is publicly available online.

International audience

The operation and maintenance of the third generation (3G) mobile networks will be challenging. These networks will be strongly service driven, and this approach differs significantly from the traditional speech dominated in the second generation (2G) approach. Compared to 2G, in 3G, the mobile cells interact and interfere with each other more, they have hundreds of adjustable parameters, and they monitor and record data related to several hundreds of different variables in each cell. This paper shows that a neural network algorithm called the self-organizing map, together with a conventional clustering method like the k-means, can effectively be used to simplify and focus network analysis. It is shown that these algorithms help in visualizing and grouping similarly behaving cells. Thus, it is easier for a human expert to discern different states of the network. This makes it possible to perform faster and more efficient troubleshooting and optimization of the parameters of the cells. The presented methods are applicable for different radio access network technologies.

The self-organizing map (SOM) is an efficient tool for visualization and clustering of multidimensional data. It transforms the input vectors on two-dimensional grid of prototype vectors and orders them. The ordered prototype vectors are easier to visualize and explore than the original data. Mobile networks produce a huge amount of spatiotemporal data. The data consists of parameters of base stations (BS) and quality information of calls. There are two alternatives in starting the data analysis. We can build either a general one-cell-model trained using state vectors from all cells, or a model of the network using state vectors with parameters from all mobile cells. In both methods, further analysis is needed to understand the reasons for various operational states of the entire network.

The self-organizing map (SOM) is a neural network algorithm which is especially suitable for the analysis and visualization of high-dimensional data. It maps nonlinear statistical relationships between high-dimensional input data into simple geometric relationships, usually on a two-dimensional grid. The mapping roughly preserves the most important topological and metric relationships of the original data elements and, thus, inherently clusters the data. The need for visualization and clustering occurs in various engineering applications, in the analysis of complex processes or systems. In addition, SOM allows easy data fusion enabling visualization and analysis of large databases of industrial systems. As a case study, the SOM has been used to cluster the pulp and paper mills of the world.

The self-organizing map (SOM) method is a new, powerful software tool for the visualization of high-dimensional data. It converts complex, nonlinear statistical relationships between high-dimensional data into simple geometric relationships on a low-dimensional display. As it thereby compresses information while preserving the most important topological and metric relationships of the primary data elements on the display, it may also be thought to produce some kind of abstractions. The term self-organizing map signifies a class of mappings defined by error-theoretic considerations. In practice they result in certain unsupervised, competitive learning processes, computed by simple-looking SOM algorithms. Many industries have found the SOM-based software tools useful. The most important property of the SOM, orderliness of the input-output mapping, can be utilized for many tasks: reduction of the amount of training data, speeding up learning nonlinear interpolation and extrapolation, generalization, and effective compression of information for its transmission.

An adaptive algorithm combining traditional linear equalization techniques and a self-organizing neural learning algorithm is presented. The results show that the performance of the neural equalizer is insensitive to nonlinear learning distortions in dynamic discrete-signal detection. Stabilization of the self-organizing map during undistorted transmission has to be further considered to decrease the absolute mean-square error (MSE) rate of the neural equalizer. The error is due to oscillations in the self-organizing map, mainly caused by the neighborhood learning. The oscillations can be decreased by taking more samples to the map before adapting themivalues and by decreasing the neighborhood learning parameter β

The concept of data flow computing is applied to digital signal processing (DSP). A data flow signal processor (DFSP) architecture is presented. The principles of data flow computing are carefully considered in order to conform with the special properties of DSP. The bus oriented architecture is easily configured to meet various performance requirements. The DFSP architecture is most suitable for nonrecursive algorithms. Typical tasks of this nature are transforms and FIR filters. A simulation model of the DFSP architecture has been developed. Simulation results of two application examples are given.

A new architecture based on the data flow principle has been developed for signal processing purposes. The Data Flow Signal Processor (DFSP) is capable of highly parallel computation of block realizations of nonrecursive algorithms. The flexibility and modularity of the DFSP architecture are illustrated by two examples. The three-sensor problem shows the high performance of the DFSP in handling three 100 kHz input streams and performing correlation computations on them. The second example shows that 256 × 256 images can be processed with spatially restricted operations at video rate.

Explicit design procedures are given for maximally flat and equal ripple lowpass and highpass transfer functions of the form H(z) = (z ± 1)q/Dn(z) , (q = 1, 2 ..... n - 1). Efficient bandpass structures are obtained directly from the realization of H(z) using the Crochiere - Penfield method. The one-multiplier lattices of Gray and Markel realize the lowpass, highpass and bandpass transfer functions with n + 1 multipliers, if q = 1. If one-multiplier lattices with ε = +1 are used in the realization, the coefficient sensitivity peaks down in the passband of the filter.