This paper describes TreeGNG, a top-down unsupervised learning method that produces hierarchical classification schemes. TreeGNG is an extension to the Growing Neural Gas algorithm that maintains a time history of the learned topological mapping. TreeGNG is able to correct poor decisions made during the early phases of the construction of the tree, and provides the novel ability to influence th...

We present here a variant of B. Fritzke’s GNG (Growing Neural Gas) — Double GNG (DGNG). In each insertion step of the GNG only one new cell is inserted in the middle of the edge connecting Maximum Resource Vertex (MRV) and a MRV in its direct topological neighbourhood. But in our DGNG two new cells are inserted at the same time. Our goal is to speed up the convergence of the learning process. A...

Competitive learning is well-known method to process data. Various goals may be achieved using competitive learning such as classification or vector quantization. In this paper, we present a different insight into the principle of supervised competitive learning. An innovative approach to the supervised self-organization is suggested. The method is based on different handling of input data labe...

Growing neural gas (GNG) has been successfully applied to unsupervised learning problems. However, GNG-inspired approaches can also be applied to classification problems, provided they are extended with an appropriate labelling function. Most approaches along these lines have so far relied on strategies which label neurons a posteriori, after the training has been completed. As a consequence, s...

An incremental network model is introduced which is able to learn the important topological relations in a given set of input vectors by means of a simple Hebb-like learning rule. In contrast to previous approaches like the "neural gas" method of Martinetz and Schulten (1991, 1994), this model has no parameters which change over time and is able to continue learning, adding units and connection...

This thesis will describe the following eight prototype-based clustering algorithms: SOM, NG, ENG, MNG, GCS, GNG, RGNG, MGNG. Thereto the basic principles of the algorithms functionality are explained and the a priori defined parameters are indicated. After this explanation their field of application are presented and a flow sheet of the functional specialisation/generalisation among the given ...

Growing Neural Gas (GNG) algorithm is an unsupervised learning which belongs to the competitive family. Since then, GNG has been a subject vaious developments and implementations found in literatures for two main reasons: first, number of neurons (i.e., nodes) adaptive. Meaning, it periodically changed through adding new removing old accordingly order find best network captures topological stru...

Incremental artificial neural networks grow when they learn and shrink when they forget. Competitive Hebbian learning generates the network structure by addition and removal of cells and links. Thus, no network design phase is necessary. The growing cell structure and the growing neural gas network may replace common feed-forward networks in a lot of classification and interpolation tasks.

Web Usage Mining becomes a vital aspect in network traffic analysis. Previous study on Web usage mining using a synchronized Clustering, Neural based approach has shown that the usage trend analysis very much depends on the performance of the clustering of the number of requests. Self Organizing Networks is useful for representation of building unsupervised learning, clustering, and Visualizati...