Counterpropagation networks applied to the classification of alkanes through infrared spectra

Mackenzie, Mathew David

doi:10.1007/bf01414354

Cited by 3 publications

(3 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This has two main advantages: firstly, performance becomes independent of initialisation; and secondly, an unusually large upper learning coefficient bound may be used, resulting in very rapid training. CDUL was developed originally as a means to overcome many of the training difficulties associated with unsupervised learning, the most obvious being that an exemplar may have resulted from training vectors of more than one class, and would hence be positioned poorly in the feature space [11]. This problem is associated to the difficulty of representing different classes of data, which exist within overlapping or disjoint regions of the feature space.…”

Section: Class Directed Unsupervised Learning and Related Networkmentioning

confidence: 99%

See 1 more Smart Citation

Enhancements to the dynamic self-organising neural network

Mackenzie¹

1996

Neural Comput & Applic

Self Cite

View full text Add to dashboard Cite

Class Directed Unsupervised Learning (CD UL) [7,8], having an input layer and a fully connected one-dimensional Kohonen layer which is used to conceptually separate data classes. These networks, along with CDUL, have all been developed from the Kohonen Self-Organising Map (SOM) [9,10] and use variations of the associated Kohonen learning rule. In each of these networks, the Kohonen nodes act as exemplars within an adaptive look-up table, and take on values corresponding to the centroids of each distinct data cluster in the feature space. Where CDUL differs is that the input layer is extended to include a class description of each training vector (see Fig. 1), and it is this additional information that provides CDUL with its interesting properties.The reason for the success of CDUL stems from two factors: firstly, the inclusion of class information as an integral section of the input vector, can be used to prevent Kohonen nodes from being trained by vectors of more than one class; and secondly, the size of the Kohonen layer can be computed

show abstract

Section: Class Directed Unsupervised Learning and Related Networkmentioning

confidence: 99%

“…12. Find winning node: select the minimum distance of those computed in (11) and designate the node at this distance from x to be k**.…”

Section: Appendix: Cdul2 Algorithm With and Without Fastcdul Without mentioning

confidence: 99%

Enhancements to the dynamic self-organising neural network

Mackenzie¹

1996

Neural Comput & Applic

Self Cite

View full text Add to dashboard Cite

show abstract

“…Previous studies using uniflow counter-propagation networks as chemical classification systems given infrared absorption spectra as input [18] have shown that, although good results may be obtained through a trial and error process of altering network parameters prior to training, the unsupervised Kohonen layer training does suffer from two main shortcomings, especially if classes within the training data are indistinct, i.e. overlapping or disjoint 9 First, it is possible that a single Kohonen node may be trained by vectors of more than one class 9 If this occurs, then the final positioning of the Kohonen node in feature space will not be representative of any single class of data, and hence cause error at the recall (or testing) stage of network operation.…”

Section: Introductionmentioning

confidence: 99%

CDUL: Class directed unsupervised learning

Mackenzie¹

1995

Neural Comput & Applic

View full text Add to dashboard Cite

A novel neural network called Class Directed Unsupervised Learning (CDUL) is introduced. The architecture, based on a Kohonen self-organising network, uses additional input nodes to feed class knowledge to the network during training, in order to optimise the final positioning of Kohonen nodes in feature space. The structure and training of CDUL networks is detailed, showing that (a) networks cannot suffer from the problem of single Kohonen nodes being trained by vectors of more than one class, (b) the number of Kohonen nodes necessary to represent the classes is found during training, and (c) the number of training set passes CD UL requires is low in comparison to similar networks. CDUL is subsequently applied to the classification of chemical excipients from Near Infrared (NIR) reflectance spectra, and its performance compared with three other unsupervised paradigms. The results thereby obtained demonstrate a superior performance which remains relatively constant through a wide range of network parameters.

show abstract

Real-time Monitoring of Environmental Pollutants in the Workplace Using Neural Networks and FTIR Spectroscopy

Cartwright

Porter

2003

Soft Computing Approaches in Chemistry

View full text Add to dashboard Cite

Counterpropagation networks applied to the classification of alkanes through infrared spectra

Cited by 3 publications

References 9 publications

Enhancements to the dynamic self-organising neural network

Enhancements to the dynamic self-organising neural network

CDUL: Class directed unsupervised learning

Real-time Monitoring of Environmental Pollutants in the Workplace Using Neural Networks and FTIR Spectroscopy

Contact Info

Product

Resources

About