Kohonen's Neural Network Algorithm Research Articles

Classification and spectrum optimization method of grease based on infrared spectrum

AbstractThe infrared (IR) absorption spectral data of 63 kinds of lubricating greases containing six different types of thickeners were obtained using the IR spectroscopy. The Kohonen neural network algorithm was used to identify the type of the lubricating grease. The results show that this machine learning method can effectively eliminate the interference fringes in the IR spectrum, and complete the feature selection and dimensionality reduction of the high-dimensional spectral data. The 63 kinds of greases exhibit spatial clustering under certain IR spectrum recognition spectral bands, which are linked to characteristic peaks of lubricating greases and improve the recognition accuracy of these greases. The model achieved recognition accuracy of 100.00%, 96.08%, 94.87%, 100.00%, and 87.50% for polyurea grease, calcium sulfonate composite grease, aluminum (Al)-based grease, bentonite grease, and lithium-based grease, respectively. Based on the different IR absorption spectrum bands produced by each kind of lubricating grease, the three-dimensional spatial distribution map of the lubricating grease drawn also verifies the accuracy of classification while recognizing the accuracy. This paper demonstrates fast recognition speed and high accuracy, proving that the Kohonen neural network algorithm has an efficient recognition ability for identifying the types of the lubricating grease.

Identification of homogenous regions in rain-fed watershed using Kohonen neural networks

Identification of homogenous regions and their ranking is important to formulate appropriate strategies for suitable conservation and management practices within the watershed. In this study, Kohonen neural network (KNN) is employed to classify the micro-watersheds of Kaddam watershed in middle Godavari basin (India) into homogeneous groups. KNN algorithm learns to cluster groups of similar input patterns from a high-dimensional input space in a non-linear fashion onto a low-dimensional layer of neurons. Ten geo-morphological parameters are used for classification of micro-watersheds. An optimal number of groups is chosen based on two cluster validation measures, the Davies–Bouldin Index and the Dunn’s Index. By using the KNN method, 18 micro-watersheds are grouped into five homogenous groups based on selected watershed parameters. The clustering results showed concurrence with general ranking of micro-watersheds. Further, the KNN micro-watershed classification results are validated by comparing with the results of K-means clustering algorithm (KCA). From the comparative analysis, it is observed that both the algorithms give approximately similar kinds of classification of micro-watersheds. The obtained results help in identifying the groups of micro-watersheds that should be given top priority (i.e. those that require immediate conservation measures). The study suggests that identifying homogeneous regions can be helpful for effective planning and management of watersheds, and KNN can be applied effectively for micro-watershed zonation.

APPLYING OF NEURAL NETWORK ON THE WINE SENSORIAL ANALYSIS FROM BARBADOS CHERRY

ABSTRACT This paper presents an approach of artificial neural networks to predict the sensorial qualities of wines. A Kohonen network has been used as a software tool in order to increase the human skills in this kind of application. An initial prototype was implemented using the artificial neural networks technology, together with the Visual Basic software from Microsoft, for evaluation of sensorial qualities of seven samples of Barbados cherry (Malpighia glabra L.) wine. Fifty consumers chosen, perhaps, had been used to obtain the sensorial data using a hedonic scale of 1–9 times. Sensorial values of flavor, aroma and appearance obtained of the hedonic dating were compared. The characteristics of wine, such as flavor and color, were similar to the Barbados cherry fresh fruit. The consumers evaluated the wines as very good; all sensorial qualities were more than 5 in hedonic scale. Results showed that Kohonen network classified the Barbados cherry wines in a distinct group, for frequency among their sensorial responses. Kohonen network results were similar or better than statistical classification; this shows that the use of Kohonen algorithm in the sensorial analysis of wines is valid. Kohonenalgorithm is very good in clustering of Barbados cherry wine samples, and its uses in sensorial analyses of wines is promising.PRACTICAL APPLICATIONSThe results of this work have practical uses in fruit wine production of Barbados cherry and in sample selections by sensorial analysis. Thus, it shows the optimal conditions to produce the Barbados cherry wine, and the effects of fruit pulp mass and sugar content on the sensorial qualities of wines had been evaluated. A visual basic program had been elaborated, in this work, by Kohonen neural network algorithm, which will be a powerful tool for taking of decision in sensorial analyses of beverages, foods and too much products that use the sensorial methods to make the comparison between them.

Cluster-based and association-based visualization systems as information exploration tools

The main purpose of this study is to investigate whether two of the algorithms, variance of Ward's hierarchical clustering algorithm and a Kohonen neural network algorithm, can help improve information exploration of unknown data collections. The initial results of the study indicate that both BiblioMapper-based and Kohonen SOM-based algorithms can successfully categorize heterogeneous data collections into manageable sub-spaces that users can successfully navigate to locate a document of interest. Both BiblioMapper and Kohonen SOM worked best with browsing tasks that were very broad, in which subjects skipped around between categories. Subjects who preferred keyword search and those who wanted to use the more familiar mental models (alphabetical organization) for browsing found that the map did not work well.

Image compression by self-organized Kohonen map

This paper presents a compression scheme for digital still images, by using the Kohonen's neural network algorithm, not only for its vector quantization feature, but also for its topological property. This property allows an increase of about 80% for the compression rate. Compared to the JPEG standard, this compression scheme shows better performances (in terms of PSNR) for compression rates higher than 30.