Traditional Self-organizing Map Research Articles

Biologically Inspired Self-Organizing Computational Model to Mimic Infant Learning

Recent technological advancements have fostered human–robot coexistence in work and residential environments. The assistive robot must exhibit humane behavior and consistent care to become an integral part of the human habitat. Furthermore, the robot requires an adaptive unsupervised learning model to explore unfamiliar conditions and collaborate seamlessly. This paper introduces variants of the growing hierarchical self-organizing map (GHSOM)-based computational models for assistive robots, which constructs knowledge from unsupervised exploration-based learning. Traditional self-organizing map (SOM) algorithms have shortcomings, including finite neuron structure, user-defined parameters, and non-hierarchical adaptive architecture. The proposed models overcome these limitations and dynamically grow to form problem-dependent hierarchical feature clusters, thereby allowing associative learning and symbol grounding. Infants can learn from their surroundings through exploration and experience, developing new neuronal connections as they learn. They can also apply their prior knowledge to solve unfamiliar problems. With infant-like emergent behavior, the presented models can operate on different problems without modifications, producing new patterns not present in the input vectors and allowing interactive result visualization. The proposed models are applied to the color, handwritten digits clustering, finger identification, and image classification problems to evaluate their adaptiveness and infant-like knowledge building. The results show that the proposed models are the preferred generalized models for assistive robots.

Maximum correntropy self-organizing map for waveform classification

The self-organizing map (SOM) is a well-known unsupervised classification method for seismic facies analysis. The conventional SOM networks and their mutation methods are mostly based on the Euclidean distance measurement, which estimates the global similarity between the input data and weights of neuron. Such measurement is suitable for signals with a Gaussian distribution but may not work well for seismic data for which the dominant structural features (e.g., stratigraphic layers, faults, channels, and salt bodies) typically appear some specially distributed patterns. Therefore, we propose to modify the SOM by introducing a new distance measurement criterion based on the maximum correntropy. The correntropy is mainly used to measure the local similarity between variables, especially for non-Gaussian signals, and therefore is more suitable than the Euclidean distance to adaptively characterize the seismic structural similarities for seismic classification tasks. Validations on two synthetic data sets indicate that the proposed method yields more accurate classification results and is more robust to noise than the traditional SOM algorithm. Comparison experiments on a field data example also demonstrate that the proposed method can delineate the distribution of seismic facies more effectively and stably than the traditional one.

Mapas auto-organizáveis de Kohonen no estudo da dissimilaridade genética entre cultivares e genótipos de soja

Abstract The objective of this work was to evaluate the genetic dissimilarity between soybean cultivars and genotypes for the selection of parents, as well as to propose a new method for using Kohonen’s self-organizing maps (SOMs) and to test its efficiency through Anderson’s discriminant analysis. The morphoagronomic descriptors of soybean cultivars and genotypes were evaluated. For data analysis, SOM-type artificial neural networks were used. The proposed method allowed the determination of the best network architecture in a nonsubjective way. Furthermore, at the beginning of training, it was possible to mitigate the randomness effect of the synaptic weights on the formed clusters. Six dissimilar clusters were formed; therefore, there is genetic dissimilarity between soybean cultivars and genotypes. Cultivars C25, C8, and C13 can be combined with C36, C31, C32, and C33 because they show good yield-related attributes and high dissimilarity. The proposed methodology is advantageous in comparison with the use of traditional SOMs, besides being efficient due to clustering consistency according to Anderson’s discriminant analysis.

Modified weight initialization in the self-organizing map using Nguyen-Widrow initialization algorithm

Self-organizing map (SOM) is a neural network trained using unsupervised learning for clustering. The clustering result and learning speed of SOM, however, is dependent on the initial weights which are randomly initialized with a low close to zero value from the range of vectors within the given input space data negatively affecting training speed and clustering result. The study used the Nguyen-Widrow initialization algorithm to initialize the weights of SOM and speed up the training process. The cluster error rate and the number of iterations to converge to final clustering is recorded and compared with the traditional SOM to determine the performance of the modified SOM. The result reveals that the modified SOM algorithm produces better cluster results and improved training speed as compared to traditional SOM. Hence, Nguyen-Widrow algorithm for initialization of weights in the SOM yields better cluster result and improved training speed of the algorithm in terms of the number of iterations.

Position Estimation Based on Grid Cells and Self-Growing Self-Organizing Map.

As the basis of animals' natal homing behavior, path integration can continuously provide current position information relative to the initial position. Some neurons in freely moving animals' brains can encode current positions and surrounding environments by special firing patterns. Research studies show that neurons such as grid cells (GCs) in the hippocampus of animals' brains are related to the path integration. They might encode the coordinate of the animal's current position in the same way as the residue number system (RNS) which is based on the Chinese remainder theorem (CRT). Hence, in order to provide vehicles a bionic position estimation method, we propose a model to decode the GCs' encoding information based on the improved traditional self-organizing map (SOM), and this model makes full use of GCs' firing characteristics. The details of the model are discussed in this paper. Besides, the model is realized by computer simulation, and its performance is analyzed under different conditions. Simulation results indicate that the proposed position estimation model is effective and stable.

Constraining self-organizing map facies analysis with stratigraphy: An approach to increase the credibility in automatic seismic facies classification

Pattern recognition-based seismic facies analysis techniques are commonly used in modern quantitative seismic interpretation. However, interpreters often treat techniques such as artificial neural networks and self-organizing maps (SOMs) as a “black box” that somehow correlates a suite of attributes to a desired geomorphological or geomechanical facies. Even when the statistical correlations are good, the inability to explain such correlations through principles of geology or physics results in suspicion of the results. The most common multiattribute facies analysis begins by correlating a suite of candidate attributes to a desired output, keeping those that correlate best for subsequent analysis. The analysis then takes place in attribute space rather than ([Formula: see text], [Formula: see text], and [Formula: see text]) space, removing spatial trends often observed by interpreters. We add a stratigraphy layering component to a SOM model that attempts to preserve the intersample relation along the vertical axis. Specifically, we use a mode decomposition algorithm to capture the sedimentary cycle pattern as an “attribute.” If we correlate this attribute to the training data, it will favor SOM facies maps that follow stratigraphy. We apply this workflow to a Barnett Shale data set and find that the constrained SOM facies map shows layers that are easily overlooked on traditional unconstrained SOM facies map.

Automatic Root Cause Analysis for LTE Networks Based on Unsupervised Techniques

The increase in the size and complexity of current cellular networks is complicating their operation and maintenance tasks. While the end-to-end user experience in terms of throughput and latency has been significantly improved, cellular networks have also become more prone to failures. In this context, mobile operators start to concentrate their efforts on creating self-healing networks, i.e., those networks capable of troubleshooting in an automatic way, making the network more reliable and reducing costs. In this paper, an automatic diagnosis system based on unsupervised techniques for Long-Term Evolution (LTE) networks is proposed. In particular, this system is built through an iterative process, using self-organizing maps (SOMs) and Ward's hierarchical method, to guarantee the quality of the solution. Furthermore, to obtain a number of relevant clusters and label them properly from a technical point of view, an approach based on the analysis of the statistical behavior of each cluster is proposed. Moreover, with the aim of increasing the accuracy of the system, a novel adjustment process is presented. It intends to refine the diagnosis solution provided by the traditional SOM according to the so-called silhouette index and the most similar cause on the basis of the minimum $X$ th percentile of all distances. The effectiveness of the developed diagnosis system is validated using real and simulated LTE data by analyzing its performance and comparing it with reference mechanisms.

Multi-class ERP-based BCI data analysis using a discriminant space self-organizing map.

Emotional or non-emotional image stimulus is recently applied to event-related potential (ERP) based brain computer interfaces (BCI). Though the classification performance is over 80% in a single trial, a discrimination between those ERPs has not been considered. In this research we tried to clarify the discriminability of four-class ERP-based BCI target data elicited by desk, seal, spider images and letter intensifications. A conventional self organizing map (SOM) and newly proposed discriminant space SOM (ds-SOM) were applied, then the discriminabilites were visualized. We also classify all pairs of those ERPs by stepwise linear discriminant analysis (SWLDA) and verify the visualization of discriminabilities. As a result, the ds-SOM showed understandable visualization of the data with a shorter computational time than the traditional SOM. We also confirmed the clear boundary between the letter cluster and the other clusters. The result was coherent with the classification performances by SWLDA. The method might be helpful not only for developing a new BCI paradigm, but also for the big data analysis.

Comparison of visualization of optimal clustering using self-organizing map and growing hierarchical self-organizing map in cellular manufacturing system

The present research deals with the cell formation problem (CFP) of cellular manufacturing system which is a NP-hard problem thus, the development of optimum machine-part cell formation algorithms has always been the primary attraction in the design of cellular manufacturing system. In this proposed work, the self-organizing map (SOM) approach has been used which is able to project data from a high-dimensional space to a low-dimensional space so it is considered a visualized approach for explaining a complicated CFP data set. However, for a large data set with a high dimensionality, a traditional flat SOM seems difficult to further explain the concepts inside the clusters. We propose one such possible solution for a large CFP data set by using the SOM in a hierarchical manner known as growing hierarchical self-organizing map (GHSOM). In the present work, the two novel contributions using GHSOM are: the choice of optimum architecture through the minimum pattern units extracted at layer 1 for the respective threshold values and selection. Furthermore, the experimental results clearly indicated that the machine-part visual clustering using GHSOM can be successfully applied in identifying a cohesive set of part family that is processed by a machine group. Computational experience specifically with the proposed GHSOM algorithm, on a set of 15 CFP problems from the literature, has shown that it performs remarkably well. The GHSOM algorithm obtained solutions that are at least as good as the ones found the literature. For 75% of the cell formation problems, the GHSOM algorithm improved the goodness of cell formation through GTE performance measure using SOM as well as best one from the literature, in some cases by as much as more than 12.81% (GTE). Thus, comparing the results of the experiment in this paper with the SOM and GHSOM using the paired t-test it has been revealed that the GHSOM approach performed better than the SOM approach so far the group technology efficiency (GTE) measures of performance of the goodness of cell formation is concerned.

P1DSOM — A FAST SEARCH ALGORITHM FOR HIGH-DIMENSIONAL FEATURE SPACE PROBLEMS

The self-organizing map (SOM) is a traditional neural network algorithm used to achieve feature extraction, clustering, visualization and data exploration. However, it is known that the computational cost of the traditional SOM, used to search for the winner neuron, is expensive especially in case of treating high-dimensional data. In this paper, we propose a novel hierarchical SOM search algorithm which significantly reduces the expensive computational cost associated with traditional SOM. It is shown here that the computational cost of the proposed approach, compared to traditional SOM, to search for the winner neuron is reduced into O(D1 + D2 + ⋯ + DN) instead of O(D1 × D2 × ⋯ × DN), where Dj is the number of neurons through a dimension dj of the feature map. At the same time, the new algorithm maintains all merits and qualities of the traditional SOM. Experimental results show that the proposed algorithm is a good alternate to traditional SOM, especially, in high-dimensional feature space problems.

Adaptive training of cortical feature maps for a robot sensorimotor controller

This work investigates self-organising cortical feature maps (SOFMs) based upon the Kohonen Self-Organising Map (SOM) but implemented with spiking neural networks. In future work, the feature maps are intended as the basis for a sensorimotor controller for an autonomous humanoid robot. Traditional SOM methods require some modifications to be useful for autonomous robotic applications. Ideally the map training process should be self-regulating and not require predefined training files or the usual SOM parameter reduction schedules. It would also be desirable if the organised map had some flexibility to accommodate new information whilst preserving previous learnt patterns. Here methods are described which have been used to develop a cortical motor map training system which goes some way towards addressing these issues. The work is presented under the general term ‘Adaptive Plasticity’ and the main contribution is the development of a ‘plasticity resource’ (PR) which is modelled as a global parameter which expresses the rate of map development and is related directly to learning on the afferent (input) connections. The PR is used to control map training in place of a traditional learning rate parameter. In conjunction with the PR, random generation of inputs from a set of exemplar patterns is used rather than predefined datasets and enables maps to be trained without deciding in advance how much data is required. An added benefit of the PR is that, unlike a traditional learning rate, it can increase as well as decrease in response to the demands of the input and so allows the map to accommodate new information when the inputs are changed during training.

Implementation and Applications of Tri-State Self-Organizing Maps on FPGA

This paper introduces a tri-state logic self-organizing map (bSOM) designed and implemented on a field programmable gate array (FPGA) chip. The bSOM takes binary inputs and maintains tri-state weights. A novel training rule is presented. The bSOM is well suited to FPGA implementation, trains quicker than the original self-organizing map (SOM), and can be used in clustering and classification problems with binary input data. Two practical applications, character recognition and appearance-based object identification, are used to illustrate the performance of the implementation. The appearance-based object identification forms part of an end-to-end surveillance system implemented wholly on FPGA. In both applications, binary signatures extracted from the objects are processed by the bSOM. The system performance is compared with a traditional SOM with real-valued weights and a strictly binary weighted SOM.

Exploratory geospatial data analysis using the GeoSOM suite

Clustering constitutes one of the most popular and important tasks in data analysis. This is true for any type of data, and geographic data is no exception. In fact, in geographic knowledge discovery the aim is, more often than not, to explore and let spatial patterns surface rather than develop predictive models. The size and dimensionality of the existing and future databases stress the need for efficient and robust clustering algorithms. This need has been successfully addressed in the context of general-purpose knowledge discovery. Geographic knowledge discovery, nonetheless can still benefit from better tools, especially if these tools are able to integrate geographic information and aspatial variables in order to assist the geographic analyst’s objectives and needs. Typically, the objectives are related with finding spatial patterns based on the interaction between location and aspatial variables. When performing cluster-based analysis of geographic data, user interaction is essential to understand and explore the emerging patterns, and the lack of appropriate tools for this task hinders a lot of otherwise very good work.In this paper, we present the GeoSOM suite as a tool designed to bridge the gap between clustering and the typical geographic information science objectives and needs. The GeoSOM suite implements the GeoSOM algorithm, which changes the traditional Self-Organizing Map algorithm to explicitly take into account geographic information. We present a case study, based on census data from Lisbon, exploring the GeoSOM suite features and exemplifying its use in the context of exploratory data analysis.

Effects of Irregular Topology in Spherical Self-Organizing Maps

The authors explore the effect of different topologies on properties of self-organizing maps (SOM). They suggest several diagnostics for measuring topology-induced errors in SOM and use these in a comparison of four different topologies. The results show that SOM is less sensitive to localized irregularities in the network structure than the literature may otherwise suggest. Further, the results support the use of spherical topologies as a solution to the boundary problem in traditional SOM.

Self-organizing feature map for cluster analysis in multi-disease diagnosis

Aiming at the multi-disease diagnosis, a self-organizing map (SOM) is developed. In this paper the tomato disease features are extracted and a mapping relationship between the diseases and the features is created. The inaccurate clustering of traditional SOM algorithm is analyzed. According to the analysis, Euclidean distance is taken as the main discrimination, and the adjacent-searching algorithm is optimized. Using the optimized algorithm, the cluster results of input samples are obtained, features of diseases are mapped, and a multi-disease diagnosis model is developed. The proposed SOM-based model has two layers. The feature array of diseases can be accurately and rapidly sorted and clustered using this model. This model can achieve an accurate diagnosis of multi-diseases. The simulation results show that the proposed model performs well and the proposed multi-disease diagnosis is effective.

Parameterless-Growing-SOM and Its Application to a Voice Instruction Learning System

An improved self-organizing map (SOM), parameterless-growing-SOM (PL-G-SOM), is proposed in this paper. To overcome problems existed in traditional SOM (Kohonen, 1982), kinds of structure-growing-SOMs or parameter-adjusting-SOMs have been invented and usually separately. Here, we combine the idea of growing SOMs (Bauer and Villmann, 1997; Dittenbach et al. 2000) and a parameterless SOM (Berglund and Sitte, 2006) together to be a novel SOM named PL-G-SOM to realize additional learning, optimal neighborhood preservation, and automatic tuning of parameters. The improved SOM is applied to construct a voice instruction learning system for partner robots adopting a simple reinforcement learning algorithm. User's instructions of voices are classified by the PL-G-SOM at first, then robots choose an expected action according to a stochastic policy. The policy is adjusted by the reward/punishment given by the user of the robot. A feeling map is also designed to express learning degrees of voice instructions. Learning and additional learning experiments used instructions in multiple languages including Japanese, English, Chinese, and Malaysian confirmed the effectiveness of our proposed system.

Cluster identification and separation in the growing self-organizing map: application in protein sequence classification

Growing self-organizing map (GSOM) has been introduced as an improvement to the self-organizing map (SOM) algorithm in clustering and knowledge discovery. Unlike the traditional SOM, GSOM has a dynamic structure which allows nodes to grow reflecting the knowledge discovered from the input data as learning progresses. The spread factor parameter (SF) in GSOM can be utilized to control the spread of the map, thus giving an analyst a flexibility to examine the clusters at different granularities. Although GSOM has been applied in various areas and has been proven effective in knowledge discovery tasks, no comprehensive study has been done on the effect of the spread factor parameter value to the cluster formation and separation. Therefore, the aim of this paper is to investigate the effect of the spread factor value towards cluster separation in the GSOM. We used simple k-means algorithm as a method to identify clusters in the GSOM. By using Davies–Bouldin index, clusters formed by different values of spread factor are obtained and the resulting clusters are analyzed. In this work, we show that clusters can be more separated when the spread factor value is increased. Hierarchical clusters can then be constructed by mapping the GSOM clusters at different spread factor values.

Modeling multisensory enhancement with self-organizing maps.

Self-organization, a process by which the internal organization of a system changes without supervision, has been proposed as a possible basis for multisensory enhancement (MSE) in the superior colliculus (Anastasio and Patton, 2003). We simplify and extend these results by presenting a simulation using traditional self-organizing maps, intended to understand and simulate MSE as it may generally occur throughout the central nervous system. This simulation of MSE: (1) uses a standard unsupervised competitive learning algorithm, (2) learns from artificially generated activation levels corresponding to driven and spontaneous stimuli from separate and combined input channels, (3) uses a sigmoidal transfer function to generate quantifiable responses to separate inputs, (4) enhances the responses when those same inputs are combined, (5) obeys the inverse effectiveness principle of multisensory integration, and (6) can topographically congregate MSE in a manner similar to that seen in cortex. Thus, the model provides a useful method for evaluating and simulating the development of enhanced interactions between responses to different sensory modalities.

Classification of audio signals using Fuzzy c-Means with divergence-based Kernel

A content-based audio retrieval method based Fuzzy c-Means algorithm with divergence kernel (FCM-DK) is proposed in this paper. FCM-DK is based on the Fuzzy c-Means algorithm and employs a kernel method for data transformation. The kernel method adopted in FCM-DK is used to transform the feature data of audio signals into a feature space of a higher dimensionality so that nonlinear problems residing in the input space can be linearly solved in the feature space. In order to deal with Gaussian probability density function (GPDF) data, a divergence-based kernel employing a divergence distance measure for its similarity measure is used for data transformation. The proposed method exploits the statistical nature of the audio data to improve the classification accuracy. Experiments and results on various data sets demonstrate that the proposed classification method outperforms conventional algorithms such as the traditional self-organizing map (SOM) and the Fuzzy c-Means (FCM) 20.83% and 17.5% in terms of accuracy.

A constrained neural learning rule for eliminating the border effect in online self-organising maps

The self-organising map (SOM) is a concise and powerful algorithm for clustering and visualisation of high-dimensional data. However, this robust algorithm still suffers from the border effect. Most of the approaches proposed to eliminate this effect use a borderless topological structure. We prefer to keep the original topological structure of the SOM for visualisation. A novel approach is proposed for the elimination of the border effect from the perspective of self-organising learning. Based on an assumption that the best matching unit (BMU) should be the most active unit, the approach proposes that the BMU should move more towards its associated input sample than its neighbours in the fine-tuned learning stage. Our constrained approach emphasises the effect of the lateral connections and neutralises the effect on the distance between the input sample and units. This approach is able to make units of the map stretch wider than the traditional SOM and thus the border effect is alleviated. Our proposed approach is proved to satisfy the requirements of the topologically ordered neural networks and is evaluated by both qualitative and quantitative criteria. All experiments conclude that performance is improved if the proposed constrained learning rule is used.