Implementation Of Self-organizing Map Research Articles

Implementation of Self-Organizing Map (SOM) Algorithm for Image Classification of Medicinal Weeds

Wild plants or weeds often become enemies or disturb the main cultivated plants. In its development, wild plants or weeds actually have ingredients that are beneficial to the body and can be used as medicine. However, many people still need knowledge about the types of weed plants that have medicinal properties, especially the leaves. The purpose of this research is to classify the image of weed leaves with medicinal properties based on color and texture characteristics with an artificial neural network using a Self-Organizing Map (SOM). To improve information in feature extraction, RGB and HSV color features are used as well as texture features with Gray Level Co-occurrence Matrix (GLCM). Furthermore, the results of feature extraction will be identified as groups or classes with the Self-Organizing Map (SOM) algorithm which divides the input pattern into several groups so that the network output is in the form of a group that is most similar to the input provided. The test produces a precision value of 91.11%, a recall value of 88.17% and an accuracy value of 89.44%. The results of the accuracy of the SOM model for image classification on medicinal weed leaves are in the good category.

Implementing in-situ self-organizing maps with memristor crossbar arrays for data mining and optimization

A self-organizing map (SOM) is a powerful unsupervised learning neural network for analyzing high-dimensional data in various applications. However, hardware implementation of SOM is challenging because of the complexity in calculating the similarities and determining neighborhoods. We experimentally demonstrated a memristor-based SOM based on Ta/TaOx/Pt 1T1R chips for the first time, which has advantages in computing speed, throughput, and energy efficiency compared with the CMOS digital counterpart, by utilizing the topological structure of the array and physical laws for computing without complicated circuits. We employed additional rows in the crossbar arrays and identified the best matching units by directly calculating the similarities between the input vectors and the weight matrix in the hardware. Using the memristor-based SOM, we demonstrated data clustering, image processing and solved the traveling salesman problem with much-improved energy efficiency and computing throughput. The physical implementation of SOM in memristor crossbar arrays extends the capability of memristor-based neuromorphic computing systems in machine learning and artificial intelligence.

Implementation of SOM (Self Organizing Maps) for Identification of Tomato Fruit Maturity

Tomat salah satu buah sekaligus sayuran dengan harga yang ekonimis dan sering kali kita jumpai. Harganya yang tidak terlalu mahal dan banyak mengandung vitamin dan mineral serta antioksidan yang tinggi, membuat tomat banyak dibeli dimasyarakat. Sehingga membuat petani buah tomat untuk terus meningkatkan mutu dan kualitas terhadap pelayanan kepada kosumen. Tetapi kadang banyak petani yang menjual buah tomatnya dengan kematangan yang tidak seragam, sehingga kadang yang sudah matang tertindih oleh yang mengkal atau mentah, sehingga banyak yang membusuk. Dari permasalahan ini, maka penulis ingin membuah sebuah aplikasi untuk memilah kematangan buah tomat mentah, mengkal dan matang. Penggunaan metode Self Organizing Map (SOM) ini bertujuan untuk melakukan pengelompokkan antara tomat yang mentah, mengkal dan matang. SOM tidak akan menghentikan proses iterasinya selama jumlah iterasinya belum mencapai target yang diharapkan. Hasil penelitian yang dihasilkan adalah sebuah aplikasi identifikasi kematangan buah tomat untuk mempermudah petani atau masyarakat luas dengan cepat. Implementasi pengujian identifikasi kematangan buah tomat ini adalah 91,11%.

Design space exploration and optimization using self-organizing maps

Identifying regions of interest (RoI) in the design space is extremely useful while building metamodels with limited computational budget. Self-organizing maps (SOM) are used as a visualization technique for design space exploration that permits identifying RoI. Conventional implementation of SOM is susceptible to folds or intersections that hinder visualizing the design space. This work proposes a modified SOM algorithm whose maps are interpretable and that does not fold and allows smoother input and performance space visualization. The modified algorithm enables identification of RoI and additional sampling in the identified RoI allows building accurate Kriging metamodel, which is then used for optimization. The proposed approach is demonstrated on benchmark nonlinear analytical examples and two practical engineering design examples. Results show that the proposed approach is highly efficient in identifying the RoI and in obtaining the optima with less samples.

Hardware Implementation of Self-Organizing Map

Hardware Implementation of Self-Organizing Map

OFDM symbol identification by an unsupervised learning system under dynamically changing channel effects

Orthogonal frequency-division multiplexing (OFDM) is one of the most successful digital communication techniques. Nevertheless, the decrease in inter-symbol interference in quadrature amplitude modulation (QAM) over dispersive channels is still challenging. Different researches recently proposed the idea of using unsupervised learning as an alternative to the classic approaches to equalization of OFDM channels. In those purposes, the identification of a received QAM symbol is possible by the comparison of its position on the in-phase/quadrature (IQ) plane relative to the positions of previously arrived symbols, generally processed by the Kohonen’s Self-Organizing Map (SOM) algorithm. This work presents the SOM unsupervised learning method executed on an embedded system applied to QAM symbols identification. The system is implemented on an FPGA, a configurable digital circuit able to meet the low power and parallel process requirements of mobile applications. Also, in order to extend the classical set of experiments to evaluate our system, this paper proposes a theoretical model of the time-varying scheme representing the transition between different channel characteristics, obtained from real measurements available on a public repository. The model is employed to verify our purpose under dynamically both changing and realistic conditions. On the assumption that it is provided enough IQ symbols for the initial training process, the hardware implementation of SOM is able to track and identify the time-varying distorted QAM constellation. No knowledge of channel characteristics is necessary. The system spends only some microseconds at start-up to reach about 100% performance, and no dedicated training phase is needed afterward.

Implementation of Self Organizing Map (SOM) as decision support: Indonesian telematics services MSMEs empowerment

Information technology and communication (telematics) is one of the most rapidly developing business sectors in Indonesia. It has strategic position in its contribution towards planning and implementation of developmental, economics, social, politics and defence strategies in business, communication and education. Aid absorption for the national telecommunication SMEs is relatively low; therefore, improvement is needed using analysis on business support cluster of which basis is types of business. In the study, the business support cluster analysis is specifically implemented for Indonesian telecommunication service. The data for the business are obtained from the National Census of Economic (Susenas 2006). The method used to develop cluster model is an Artificial Neural Network (ANN) system called Self-Organizing Maps (SOM) algorithm. Based on Index of Davies Bouldin (IDB), the accuracy level of the cluster model is 0.37 or can be categorized as good. The cluster model is developed to find out telecommunication business clusters that has influence towards the national economy so that it is easier for the government to supervise telecommunication business.

Urban habitats and the injury landscape

Multi-dimensional data analysis has rarely been performed on injury data. However, the implementation of sound policies can only be considered over geographical space when typologies of health vectors are considered at a global scale, and may be integrated at local level for decision making. In this study, frequency data of six injury types in Extended Golden Horseshoe area are collected and self-organizing maps (SOM) are used to construct an injury cluster system for the region. The results indicate that the variables are positively correlated and several outliers are found in the study area. The implementation of SOM brings an efficient tool for planning and assessment of spatially-explicit clusters for decision making at local and regional level. Furthermore, the visualization of multi-dimensional data allows for an integrative system of understanding the connectivity of injuries to existing land use types. The results also suggest that the injury prevention strategies should be reinforced in smaller CMAs like St. Catherine's – Niagara and Oshawa. But also, that land use typologies exist within metropolitan areas that create more probable injury outcomes. Overall our paper suggests the existence of clear clusters over geographical space that transcend traditional analysis techniques that may be adapted to ubiquitous concerns in epidemiology linked to land use planning and safer environments. Self-organizing maps may have a profound and significant role in merging decision-makers with adequate spatial information for health determinants and assessments suggesting a new integrative method of multi-dimension analysis of land use and injury data.

The Erythemato-Squamous Dermatology Diseases Severity Determination using Self-Organizing Map

A new approach based on the implementation of Self Organizing Map is presented for automated detection of erythemato-squamous diseases. The purpose of clustering techniques is in order to determinate the severity of erythemato-squamous dermatology diseases. The studied domain contained records of patients with known diagnosis. Self-Organizing Map algorithm's task was to classify the data points, in this case the patients with attribute data, to one of the six clusters (psoriasis, seboreic dermatitis, lichen planus, pityriasis rosea, cronic dermatitis, dan pityriasis rubra pilaris). The algorithm was used to detect the six erythemato-squamous diseases when 33 features defining five disease indications were used. The purpose is to determine an optimum classification scheme for this problem. The present research demonstrated that the features well represent the erythemato-squamous diseases and SOM algorithm's task achieved high classification accuracies. The best accuration for psoriasis 85,94%, seboreic dermatitis 40,48%, lichen planus 56,25%, and pityriasis rosea 82,61%, with learning rate value were 0,1, 0,2, 0,9, and 0,4

Implementation of Self-Organizing Maps (SOM) to analyses of environmental parameters and phytoplankton biomass in a macrotidal estuary and artificial lake

Self-Organizing Maps (SOM) have been used for patterning and visualizing ten environmental parameters and phytoplankton biomass in a mactrotidal (>10 m) Gyeonggi Bay and artificial Shihwa Lake during 1986–2004. SOM segregated study areas into four groups and ten subgroups. Two strikingly alternative states are frequently observed: the first is a diverse non-eutrophic state designated by three groups (SOM 1–3), and the second is a eutrophic state (SOM 4: Shihwa Lake and Upper Gyeonggi Bay; summer season) characterized by enhanced nutrients (3 mg l−1 dissolved inorganic nitrogen, 0.1 mg l−1 PO4) that act as a signal and response to that signal as algal blooms (24 µg chlorophyll-a l−1). Bloom potential in response to nitrification is affiliated with high temperature (r = 0.26), low salinity (r = −0.40) and suspended solids (r = –0.27). Moreover, strong stratification in the Shihwa Lake has accelerated harmful algal blooms and hypoxia. The non-eutrophic states (SOM 1–3) are characterized by macro-tidal estuaries exhibiting a tolerance to pollution with nitrogen-containing nutrients and retarding any tendency toward stratification. SOM 1 (winter) is more distinct from SOM 4 due to higher suspended solids (>50 mg l−1) caused by resuspension that induces light limitation and low chlorophyll-a (<5 µg l−1). In addition, eutrophication-induced shifts in phytoplankton communities are noticed during all the seasons in Gyeonggi Bay. Overall, SOM showed high performance for visualization and abstraction of ecological data and could serve as an efficient ecological map that can specify blooming regions and provide a comprehensive view on the eutrophication process in a macrotidal estuary.

Scalability of Self-organizing Maps on a GPU cluster using OpenCL and CUDA

We evaluate a novel implementation of a Self-Organizing Map (SOM) on a Graphics Processing Unit (GPU) cluster. Using various combinations of OpenCL, CUDA, and two different graphics cards, we demonstrate the scalability of the SOM implementation on one to eight GPUs. Results indicate that while the algorithm scales well with the number of training samples and the map size, the benefits from using the data-parallel approaches offered by the GPU are severely limited when combined with the Message Passing Interface (MPI) in this setting, and comparable to speedups of GPU-based implementations as compared to optimized sequential code. Speedups achieved range from 3 to 32, for various map and training data sizes. We also observed a performance penalty for the OpenCL implementation as compared to CUDA.

Static Performance of Wireless Localization Algorithm Exploiting Self-Organizing Maps

In wireless communications, determining the physical location of nodes (localization) is very important for many network services and protocols. This paper evaluates the static performance of wireless localization algorithm exploiting Self-Organizing Maps (SOM) to deal with this issue. Our proposed algorithm utilizes only connectivity information and information from some heard anchors in the network to determine the location of nodes. By introducing an efficient two-hop utilization scheme and the multi-hop anchor update, the algorithm has maximized the correlation between neighboring nodes and the global topology in distributed implementation of SOM. From our intensive simulations on various static network deployment scenarios, the results show that the proposed scheme achieves very good localization accuracy. It also reduces the SOM learning steps to around 15 to 30 steps to overcome the huge computational problem of the classical SOM.

Distributed Range-Free Localization Algorithm Based on Self-Organizing Maps

In Mobile Ad Hoc Networks (MANETs), determining the physical location of nodes (localization) is very important for many network services and protocols. This paper proposes a new Distributed Range-Free Localization Algorithm Based on Self-Organizing Maps (SOMs) to deal with this issue. Our proposed algorithm utilizes only connectivity information to determine the location of nodes. By utilizing the intersection areas between radio coverage of neighboring nodes, the algorithm has maximized the correlation between neighboring nodes in distributed implementation of SOM and reduced the SOM learning time. An implementation of the algorithm on Network Simulator 2 (NS-2) was done with the mobility consideration to verify the performance of the proposed algorithm. From our intensive simulations, the results show that the proposed scheme achieves very good accuracy in most cases.

Using improved self-organizing map for partial discharge diagnosis of large turbogenerators

Partial discharge (PD) classification is a powerful way to predict insulation problems of the windings of rotating machinery. While online PD tests have been carried out for over 40 years, effective diagnostic methods are still under development. In this paper, a practical diagnostic method is proposed based on an improved self-organizing map (SOM) with improved clustering indications. Three feature extraction methods are employed for the SOM implementation, including Weibull analysis, statistical operators, and fractal parameters. Experimental PD data of industrial model bars are used to validate the efficiency of using SOM for PD classification. The method is applied to investigate the turbine generator analyzer (TGA) data obtained from a power plant of British Nuclear Fuels Ltd. Diagnostic results are included to demonstrate that the relationship between the new PD measurement and historical data can be visualized and more confidential diagnostic information can be provided, especially when small-size database, new class of data, and doubtful measurements are involved in a practical environment.

Parallel Implementation of Self-Organizing Map on the Partial Tree Shape Neurocomputer

A parallel mapping of self-organizing map (SOM) algorithm is presented for a partial tree shape neurocomputer (PARNEU). PARNEU is a general purpose parallel neurocomputer that is designed for soft computing applications. Practical scalability and a reconfigurable partial tree network are the main architectural features. The presented neuron parallel mapping of SOM with on-line learning illustrates a parallel winner neuron search and a coordinate transfer that are performed in the partial tree network. Phase times are measured to analyse speedup and scalability of the mapping. The performance of the learning phase in SOM with a four processor PARNEU configuration is about 26 MCUPS and the recall phase performs 30 MCPS. Compared to other mappings done for general purpose neurocomputers, PARNEU's performance is very good.