Fuzzy Kohonen Clustering Network Research Articles

Spatial Analysis of Soil Fertility Using Geostatistical Techniques And Artificial Neural Networks

Information on the spatial variation of soil fertility attributes is an essential input for precision agriculture and soil management decision making. In this study, soil fertility assessment was carried out through the spatial distribution of thematic maps of individual properties and the subsequent integration into a digital mapping model of local fertility classes, as fundamental bases for the implementation of fertilization and amendment plans adjusted to soil status and crop requirements. For the evaluation of fertility, a systematic surface sampling was carried out in 70 sites in the "Agronomy" production field of the National University of the Central Plains "Romulo Gallegos", El Castrero sector, Juan German Roscio municipality, Guárico state, Venezuela. Ten soil variables were analyzed: pH (1:2.5), electrical conductivity (1:5), organic matter, available phosphorus, assimilable potassium, available calcium and magnesium, and the relative amounts of sand, silt and clay. Soil property maps were produced by geostatistical analysis and interpolation by ordinary kriging, and artificial intelligence techniques based on an artificial neural network classification system were applied to generate soil fertility classes using the Fuzzy Kohonen Clustering Network (FKCN) algorithm by interpolating the values of the membership function for each of the classes. The reliability of the individual maps of each soil variable was obtained by cross validation with a reliability level higher than 90%, with the exception of the variables % Clay and % Silt that presented a reliability higher than 85%. The integration of the soil attribute maps and the combination of the values of belonging to each class produced a map integrated by five soil fertility categories. The final model of digital soil fertility classes presented a reliability equivalent to 86%, which indicated a high degree of homogeneity within the soil classes obtained for fertility purposes.

Mining event logs for knowledge discovery based on adaptive efficient fuzzy Kohonen clustering network

As a digital representation of building projects, Building Information Modeling (BIM) can accumulate large volumes of log data containing hidden knowledge for deep exploration. However, such ever-increasing logs are likely to suffer from high complexity, inaccuracy, and uncertainty, which will inevitably raise challenges in uncovering latent and meaningful patterns. In order to yield satisfactory clustering quality and efficiency for better design process management, a novel clustering-based BIM event log mining approach is put forward in this paper. For one thing, a hybrid clustering algorithm named adaptive efficient fuzzy Kohonen clustering network (AEFKCN) is developed with a modified learning rate to accelerate the convergence. For another, a new clustering validity index (CVI) only relying on boundary points is designed to reduce computational complexity. An experiment is conducted in a 4 GB realistic BIM design event log dataset to validate the effectiveness of the proposed method. It begins from extracting a set of features associated with designers’ engagement and efficiency and ends up retrieving inherent insights into the person’s design behavioral patterns. Moreover, the cluster analysis can significantly distinguish the design productivity at different time periods into the high, medium, and low level, which presents a unique opportunity in understanding and assessing design productivity objectively. Practically, our method can support data-driven decision making for managers to strategically schedule personalized work for different designers, aiming to boost design efficiency and smooth the design process.

Blood vessel detection from Retinal fundas images using GIFKCN classifier

In this paper, an automated method for blood vessel detection from retinal fundus images is proposed. Initially, the method extracts the green layer of the retinal image as it contains the requisite information. The noise is removed using a noise removal filter this step will lead to the reduction of artifacts in the final results. The image features are highlighted through morphological operations i.e., top hat and bottom hat transformation on the preprocessed image. The resultant image is then clustered into two clusters representing vessel and non-vessel using generalized improved Fuzzy Kohonen Clustering Network (GIFKCN). The experiments are performed on the Digital Retinal Images for Vessel Extraction (DRIVE) database. The effectiveness of the proposed technique is analyzed using performance metrics like accuracy and sensitivity. A comparative study of the proposed technique with some of the well-known existing methods is done. The comparison shows that the proposed method has improved significantly.

BIM log mining: Exploring design productivity characteristics

Abstract A clustering-based building information modeling (BIM) log mining method is developed in this research to provide a data-driven knowledge discovery about the design productivity characteristics from a huge amount of BIM design log data. Since design behaviors are non-deterministic and subjective, a novel clustering algorithm named efficient fuzzy Kohonen clustering network (EFKCN) is utilized to produce informative clusters of different features. First, datasets are pulled out from the raw design logs and transformed into understandable forms for computers. Then, EFKCN clustering algorithm is performed in datasets at both individual and team level. Finally, analysis and prediction methods, like time analysis, regression and others, will further investigate the extracted clusters, which help managers to figure out the design preference and productivity of different designers. A case study is conducted in the real BIM design logs from an international architecture design firm with 853,520 records to illustrate the effectiveness of the proposed method. From a view of individuals, the personal design behaviors hidden in different clusters are served to arrange proper design work rationally during particular time periods. From a team perspective, the design productivity of different designers can be approximately evaluated as high, medium, and low levels in an objective and efficient manner. In the comparison of EFKCN with other popular clustering methods, EFKCN takes less time and iterations to complete the clustering process, and its clustering results achieve great compactness and separation according to the value of cluster validity indices (CVIs). Hence, this research contributes to performing the novel clustering-based BIM log mining, which acts as a powerful decision-making tool in evaluating design productivity and drawing up personalized work arrangements for a more efficient modeling process.

HEART RATE VARIABILITY CLASSIFICATION USING SADE-ELM CLASSIFIER WITH BAT FEATURE SELECTION

The electrical activity of the human heart is measured by the vital bio medical signal called ECG. This electrocardiogram is employed as a crucial source to gather the diagnostic information of a patient's cardiopathy. The monitoring function of cardiac disease is diagnosed by documenting and handling the electrocardiogram (ECG) impulses. In the recent years many research has been done and developing an enhanced method to identify the risk in the patient's body condition by processing and analysing the ECG signal. This analysis of the signal helps to find the cardiac abnormalities, arrhythmias, and many other heart problems. ECG signal is processed to detect the variability in heart rhythm; heart rate variability is calculated based on the time interval between heart beats. Heart Rate Variability HRV is measured by the variation in the beat to beat interval. The Heart rate Variability (HRV) is an essential aspect to diagnose the properties of the heart. Recent development enhances the potential with the aid of non-linear metrics in reference point with feature selection. In this paper, the fundamental elements are taken from the ECG signal for feature selection process where Bat algorithm is employed for feature selection to predict the best feature and presented to the classifier for accurate classification. The popular machine learning algorithm ELM is taken for classification, integrated with evolutionary algorithm named Self- Adaptive Differential Evolution Extreme Learning Machine SADEELM to improve the reliability of classification. It combines Effective Fuzzy Kohonen clustering network (EFKCN) to be able to increase the accuracy of the effect for HRV transmission classification. Hence, it is observed that the experiment carried out unveils that the precision is improved by the SADE-ELM method and concurrently optimizes the computation time.

Segmentation and classification models validation area mapping of peat lands as initial value of Fuzzy Kohonen Clustering Network

Ogan Komering Ilir (OKI) is located at the eastern of South Sumatra Province, 2030′-4015′ latitude and 104020′-106000′ longitude. Digital image of land was captured from Landsat 8 satellite path 124/row 062. Landsat 8 is new generation satellite which has two sensors, Operation Land Manager (OLI) and Thermal Infra-Red Sensor (TIRS). In pre-processing step, there are a geometric correction, radiometric correction, and cropping of the digital images which resulting coordinated geography. Classification uses maximum likelihood estimator algorithm. In segmentation process and classification, grey value spread is into evenly after applying histogram technique. The results of entropy value are7.42 which is the highest of result image classification, then the smallest entropy value in the result of correction mapping are 6.39. The three of them prove that they have enough high entropy value. Then the result of peatlands classification is given overall accuracy value = = 94.0012% and overall kappa value = 0.9230 so the result of classification can be considered to be right.

Expedited generation of terrain digital classes in flat areas from UAV images for precision agriculture purposes

Precision agriculture (PA) requires reasonably homogeneous areas for site-specific management. This work explores the applicability of digital terrain classes obtained from a digital elevation model derived from UAV-acquired images, to define management units in in a relative flat area of about 6 ha. Elevation, together with other terrain variables such as: slope degree, profile curvature, plan curvature, topographic wetness index, sediment transport index, were clustered using the Fuzzy Kohonen Clustering Network (FKCN). Four terrain classes were obtained. The result was compared with a map produced by a classification of soil properties previously interpolated by ordinary kriging. The results suggest that areas for site-specific management can be defined from terrain classes based on environmental covariates, saving time and cost in comparison with interpolation of soil variables.

Segmentation and Classification Models Validation Area Mapping of Peat Lands as Initial Value of Fuzzy Kohonen Clustering Network

Ogan Komering Ilir (OKI) is located at the eastern of South Sumatra Province, 2 0 30'-4 0 15' latitude and 104 0 20'-106 0 00' longitude. Digital image of land was captured from Landsat 8 satellite path 124/row 062. Landsat 8 is new generation satellite which has two sensors, Operation Land Manager (OLI) and Thermal Infra-Red Sensor (TIRS). In pre-processing step, there are a geometric correction, radiometric correction, and cropping of the digital images which resulting coordinated geography. Classification uses maximum likelihood estimator algorithm. In segmentation process and classification, grey value spread is into evenly after applying histogram technique. The results of entropy value are7.42 which is the highest of result image classification, then the smallest entropy value in the result of correction mapping are 6.39. The three of them prove that they have enough high entropy value. Then the result of peatlands classification is given overall accuracy value = = 94.0012% and overall kappa value = 0.9230 so the result of classification can be considered to be right.

Segmentation and Classification Models Validation Area Mapping of Peat Lands as Initial Value of Fuzzy Kohonen Clustering Network

Segmentation and Classification Models Validation Area Mapping of Peat Lands as Initial Value of Fuzzy Kohonen Clustering Network

Detection of 2011 Sikkim earthquake-induced landslides using neuro-fuzzy classifier and digital elevation model

In this paper, a semiautomatic method for landslide detection from satellite images and digital terrain information using generalized improved fuzzy Kohonen clustering network (GIFKCN) classifier is presented. The proposed method classifies the pre- and post-landslide images using the GIFKCN classifier which is trained using spectral indices such as normalized difference vegetation index, normalized difference building index and normalized difference water index. The changes in the vegetation class are identified using the pre- and post-classified images. Generally, landslides result in loss of vegetation; thus, using this property, candidate landslides are identified. Finally, false positives are removed using a rule set created from DEM derivatives slope and aspect. The proposed method is applied on Landsat 5 and Advanced Land Imager EO-1 satellite images to detect earthquake-induced landslides that occurred in Sikkim state of India due to the September 18, 2011, earthquake of magnitude M w = 6.9. The terrain information used is ASTER Global Digital Elevation Model of the area. The accuracy assessment of the method is done, and the results show that the landslides are identified and classified efficiently.

Performance evaluation of FMIG clustering using fuzzy validity indexes

The clustering of high-dimensional data presents a critical computational problem. Therefore, it is convenient to arrange the cluster centres on a grid with a small dimensional space that reduces computational cost and can be easily visualized. Moreover, in real applications there is no sharp boundary between classes, real datasets are naturally defined in a fuzzy context. Therefore, fuzzy clustering fits better for complex real datasets to determine the best distribution. Self-organizing map (SOM) technique is appropriate for clustering and vector quantization of high-dimensional data. In this paper we present a new fuzzy learning approach called FMIG (fuzzy multilevel interior growing self-organizing maps). The proposed algorithm is a fuzzy version of MIGSOM (multilevel interior growing self-organizing maps). The main contribution of FMIG is to define a fuzzy process of data mapping and to take into account the fuzzy aspect of high-dimensional real datasets. This new algorithm is able to auto-organize the map accordingly to the fuzzy training property of the nodes. In the second step, the introduced scheme for FMIG is clustered by means of fuzzy C-means (FCM) to discover the interior class distribution of the learned data. The validation of FCM partitions is directed by applying six validity indexes. Superiority of the new method is demonstrated by comparing it with crisp MIGSOM, GSOM (growing SOM) and FKCN (fuzzy Kohonen clustering network) techniques. Thus, our new method shows improvement in term of quantization error, topology preservation and clustering ability.

Image segmentation algorithm based on high-dimension fuzzy character and restrained clustering network

An image segmentation algorithm of the restrained fuzzy Kohonen clustering network (RFKCN) based on highdimension fuzzy character is proposed. The algorithm includes two steps. The first step is the fuzzification of pixels in which two redundant images are built by fuzzy mean value and fuzzy median value. The second step is to construct a three-dimensional (3-D) feature vector of redundant images and their original images and cluster the feature vector through RFKCN, to realize image segmentation. The proposed algorithm fully takes into account not only gray distribution information of pixels, but also relevant information and fuzzy information among neighboring pixels in constructing 3D character space. Based on the combination of competitiveness, redundancy and complementary of the information, the proposed algorithm improves the accuracy of clustering. Theoretical analyses and experimental results demonstrate that the proposed algorithm has a good segmentation performance.

Detection of 2011 Tohoku Tsunami Inundated Areas in Ishinomaki City Using Generalized Improved Fuzzy Kohonen Clustering Network

In this paper, a framework for identifying tsunami inundated areas using an innovative Generalized Improved Fuzzy Kohonen Clustering Network (GIFKCN) is proposed. GIFKCN hybridizes the Kohonen clustering network with Generalized Improved Fuzzy Partitions FCM (GIFP-FCM) algorithm to build a more efficient and effective neuro fuzzy classifier. GIFKCN classifier combines the advantages of both a neural network and fuzzy systems. A number of spectral indices are computed and the mean values of these indices are used to train the GIFKCN classifier. The novel classifier was applied to identify March 2011 Tohoku tsunami inundated areas in Ishinomaki city. The performance of the classifier is satisfactory with high overall accuracy and Kappa coefficient.

Fuzzy Kohonen clustering networks for interval data

The Fuzzy Kohonen Clustering Network combines the idea of fuzzy membership values for learning rates. It is a kind of self-organizing fuzzy neural network that can show great superiority in processing the ambiguity and the uncertainty of data sets or images. Symbolic data analysis provides suitable tools for managing aggregated data described by intervals. This paper introduces Fuzzy Kohonen Clustering Networks for partitioning interval data. The first network is based on a fixed Euclidean distance for interval and the second one considers weighted distances that change at each iteration, but are different from one cluster to another. Experiments with real and synthetic interval data sets demonstrate the usefulness of these networks.

Multi-sensor information fusion based on rough set and fuzzy Kohonen clustering network

Multi-sensor information fusion based on rough set and fuzzy Kohonen clustering network

Behavior-Based Navigation Using Heuristic Fuzzy Kohonen Clustering Network for Mobile Service Robots

This technical note presents a behavior-based navigation method for a mobile service robot with a three-wheeled omnidirectional mobile platform. A heuristic fuzzy Kohonen clustering network (FKCN) is presented to obtain laser area weights and desired robot motion heading for all types of indoor environments. An efficient behavior-based navigation structure is then proposed to merge the heuristic FKCN with three fuzzy behaviors in order to navigate the robot without any collisions in crowded and cluttered indoor environments. Simulation and experimental results are conducted to verify the feasibility and effectiveness of the proposed method.

Real-time approach for cloth simulation

This paper presents a neural technique for simulating deformable objects. For the sake of achieving real-time and reducing the computational time costs, the fuzzy Kohonen clustering network (FKCN) is used for modeling and visualization. However, determining the physical parameters of differential equation is very difficult and sometimes different equation should be solved to produce animations, which become ill-conditioned. The finite element model is normally very time-consuming, since an appropriate structure must be defined. The proposed method can be more easily implemented than the finite element method. The simulation results show that the neural network method produced plausible animation results.

Real-time monitoring of abnormal conditions based on Fuzzy Kohonen clustering network in gas metal arc welding

A real-time monitoring system based on through-the-arc sensing is developed for detecting abnormal conditions in gas metal arc welding. The transient signals of welding voltage and current during the welding process are sampled and processed by statistical analysis methods. It is found that three statistical parameters (the standard deviation, variance, and kurtosis of welding current) show obvious variations during the step disturbance, which is intentionally introduced into the T-joint test pieces by cutting a gap in the vertical plane. A Fuzzy Kohonen clustering network (FKCN) is put forward to monitor the abnormal conditions in real-time. Ten robotic welding experiments are conducted to verify the real-time monitoring system. It is found that the correct identification rate is above 90%.

Intelligent monitoring and recognition of the short-circuiting gas—metal arc welding process

This paper introduces an intelligent system for monitoring and recognition of process disturbances during short-circuiting gas-metal arc welding. It is based on the measured and statistically processed data of welding electrical parameters. A 12-dimensional array of process features is designed to describe various welding conditions and is employed as input vector of the intelligent system. Three methods, such as fuzzy c-means, neural network and fuzzy Kohonen clustering network are used to conduct process monitoring and automatic recognition. The correct recognition rates of these three methods are compared.

Mapping adaptive fuzzy Kohonen clustering network onto distributed image processing system

Algorithm of mapping adaptive fuzzy Kohonen clustering network (AFKCN) onto an image processing system with distributed memory and torus topology is presented. A hypercube is used as a structure of the AFKCN parallel program for image segmentation. The impossibility of message congestions on torus links is proved for the hypercube-to-torus XOR-embedding. Expressions for the parallel AFKCN performance analysis are given. For distributed image processing systems having two- and three-measured torus topologies the analysis shows good estimation of the parallel AFKCN implementation.