Papers by Keyword: SOM

Abstract: This paper presents a comparative study of clustering using Artificial Intelligence (AI) techniques. There are 3 methods to be compared, two methods are pure method, called Self Organising Map (SOM) which is branch of Artificial Neural Network (ANN) and Genetic Algorithm (GA), while one method is hybrid between GA and SOM, called GA-based SOM. SOM is one of the most popular method for cluster analysis. SOM will group objects based on the nearest distance between object and updateable cluster centres. However, there are disadvantages of SOM. Solution quality is depend on initial cluster centres that are generated randomly and cluster centres update algorithm is just based on a delta value without considering the searching direction. Basically, clustering case could be modelled as optimisation case. The objective function is to minimise total distance of all data to their cluster centre. Hence, GA has potentiality to be applied for clustering. Advantage of GA is it has multi searching points in finding the solution and stochastic movement from a phase to the next phase. Therefore, possibility of GA to find global optimum solution will be higher. However, there is still some possibility of GA just find near-optimum solution. The advantage of SOM is the smooth iterative procedure to improve existing cluster centres. Hybridisation of GA and SOM believed could provide better solution. In this study, there are 2 data sets used to test the performance of the three techniques. The study shows that when the solution domain is very wide then SOM and GA-based SOM perform better compared to GA while when the solution domain is not very wide then GA performs better.

Abstract: This paper describes two methods for the industrial quality inspection: Supervised classification algorithm Chi-Square Automatic Interaction Detector (CHAID) and unsupervised clustering algorithm Self-Organizing Map (SOM). The classification and clustering are modelled in IBM software SPSS. Models’ functioning is illustrated on a wheel assembly geometric features inspection. The classifying accuracies are compared for the two methods. CHAID has shown better classifying ability than SOM, while SOM can be used to improve quality of predictor values, and therefore classifiers accuracy.

Abstract: As one of the most important type of machinery, rotating machinery may malfunction due to various reasons. Sometimes the fault is a single one, but sometimes it maybe in multi-fault condition, this paper mainly focus on the latter. First, the paper gives a brief introduction of the study on multi-fault, then it introduces the mixture of Alpha stable distribution model, besides, it gives the model parameters estimation algorithm in detail, at last we use the SOM net to complete pattern recognition. The results prove that this modeling method is effective in multi-fault diagnosis in rotating machinery.

Abstract: The analysis of electromagnetic interferences (EMI) has been a heated problem in the field of Electromagnetic Compatibility (EMC). As the demand of efficiency and effectiveness is getting higher, the traditional methods have become the short board in analysis process. These methods havent provided a solution to analyze the relation among multiple EMI signals, and the data clustering and mining are currently done manually. To address this problem, in this paper we propose a one-stop method based on the wavelet packet decomposition (WPD) and self-organized feature map (SOM), aiming to provide a systematical and solution to extract and analyze multiple EMI signals. Experimental results are also provided to demonstrate the validity and efficiency of the proposed method.

Abstract: The General Status of the Refractory Metal was Reviewed Including its Characteristics and Application. the New Technology Research of Directly Prepared Refractory Metal from Refractory Metal Oxide was Discussed. FFC, OS and SOM Methods were Mainly Introduced. at the same Time, their Advantages and Disadvantages and the Difference between each other were also Pointed out. the Development Trends and Application Prospect of the Refractory Metal in Future were Prospected.

Abstract: In this paper, we present an improved text clustering algorithm. It not only maintains the self-organizing features of SOM network, but also makes up the disadvantages of the bad clustering effect caused by the inadequate selection of K-means algorithm. Firstly, data is preprocessed to form vector space model for subsequent process. Then, we analyze the features of original clustering algorithm and SOM algorithm, and plan an improved SOM clustering algorithm to overcome low stability and poor quality of original algorithm. The experimental results indicate that the improved algorithm has a higher accuracy and has a better stability, compared with the original algorithm.

Abstract: This paper presents a method of detecting pedestrians side in video frames of cluttered scenes. This detection technique is based on the idea of wavelet template and SOM neutral network. In order to make detection results more accurate and reduce computation cost, we combine background subtraction and frames difference to decide where pedestrians stand in a frame.

Abstract: An approach for establishment of soil environmental assessment model to evaluate the environmental quality level for soil environmental quality is proposed, in which the GIS and self-organizing map (SOM) techniques are integrated through investigation of soil environmental quality. In this model, self-organizing maps (SOM) and spatial interpolation were applied to cluster a concentration data set of pollutants of soil environmental quality and mapping anomaly region. An application of heavy metal concentrations in soils were surveyed to indicate the status of heavy metal contents and assess environmental quality of soils basing on spatial extraction model. The concentration of 9 metals (Cu, Pb, Zn, Cd, Ni, Cr, Hg, As and Mn) in topsoil were investigated based on samples. The samples were clustered into 3 classes by SOM. According to the concentration level of the samples, the different environmental quality levels were discriminated. The results indicate that SOM as the spatial extraction model was effective to assess the soil environmental quality.

Abstract: Recommendation systems have a prevalent cold-start problem. The problem is occurred might due to new users or new items (music) are added into the system. In this paper, the meaning of the cold-start is narrowed to that the systems do not understand the new user’s preferences. Therefore, systems can not recommend the music to users. Although many recommendation systems have a solution to reduce the cold-start, e.g., general systems utilize random to select songs. The systems random select some music works to user so that systems will know the user’s preferences after they rated the music works. However, systems may cost much time to collect the necessary information when the new user is interesting in some special types of music. Therefore, if systems select various type of music initially, the user’s preferences will be extracted more quickly. That is the cold-start problem can be reduced when the types of initial recommended music are various. In our approach, we utilize SOM to select some music from clusters. According to experiment, SOM selects type of music more average than k-means and random selection. Therefore, SOM can improve the cold-start problem and increase the precision of recommendation results.

Abstract: The Structural Health Monitoring process includes several steps like feature extraction and probabilistic decision making, which need some form of data fusion and information condensation. These take place after data acquisition and before being able to decide, if a monitored structure has faced damage. Although feature selection is an important step, the processing and suitable preparation of these data are significant, influencing the potential of decision making in various ways. With Self-Organizing Maps (SOM) a multi-purpose instrument for these tasks of pattern recognition and data interpretation is presented here. Self-Organizing Maps belong to the group of artificial neural networks and by using the special map character provide the opportunity of additional visualization. Especially when monitoring a structure over a long period of time, environmental changes often occur, which can mask the effects of damage on the dynamic behavior of the structures. As one potential application of SOM, the possibility of distinguishing between environmental changes and damage of the structure is shown. In this application a self-organizing network is trained with data of the undamaged structure and via calculation of the distance to the map a damage indicator is developed. Moreover, the distinction between different damage modes of piezoelectric sensors is presented using SOM as a tool of pattern recognition and visualization. This application uses data recorded from different damage modes extracted from one specimen of a piezoelectric element. The trained network can be compared with other piezoelectric elements mounted in a similar way to be able to detect possible sensor damage. This helps avoiding false alarms even under changing environmental conditions. Both applications have been successfully used to analyze experimental data on coupon level showing the applicability of the presented concepts.