The wavelet transform as a tool for recognition of biosignals.

The use of the wavelet transform as a signal analysis tool has been demonstrated by its successful application to the study of various signals. The first step in addressing pattern recognition problems is to define a representation that can be used for extracting the information content of signals. The sharp variation points of a signal amplitude are among the meaningful characterizations of the signal. The wavelet transform of the signal is found to be translation variant which makes it difficult for direct application in pattern recognition. However, the zero-crossings of a wavelet transform employing a particular class of wavelets can provide the translation invariant locations of the signal variation points. A zero-crossing representation augmented by the measure of the structure between the two consecutive zero-crossings has been studied by Stephane Mallat. On the basis of this representation, we demonstrate recognition of segments of biosignals embedded in streams of signals. The feasibility of employing zero-crossings of a wavelet transform as a tool in searching for a particular pattern class in the library of biosignals is explored.

An examination of inductive learning algorithms for the classification of sleep signals.

To determine their usefulness in sleep stage scoring, nine inductive learning algorithms have been tested against the sleep signals of 161 subjects representing more than 130,000 sleep events. The performance of each algorithm has been examined relative to the number of somnologist-supplied events, the simplicity of the induced rules, the percentage of all events correctly classified and the percentage of classified-events correctly classified. The last category is especially important in building reliable systems in a medical domain, where it is better for an event to be labeled "unknown" rather than incorrectly labeled. Algorithms showing the best overall performance are C4, MDL, and AIMS, generating the simplest rules, with a very high overall accuracy. PRG, a more conservative classifier, has a significantly higher accuracy on events that it is able to classify. COBWEB and the Nearest Neighbor method had marginally higher accuracy when the number of somnologist-supplied events is limited.

Computer sleep stage scoring--an expert system approach.

This paper describes the application of the expert system concept to sleep stage scoring and is aimed at managing the large variability found in such records. The primary computational burden, sleep parameter calculation, can be relieved by recognizing major classes of recordings or 'somnotypes' having internally consistent sleep parameters for each class but significant differences between classes. The initiation and evolution of somnotypes so as to maintain accuracy while introducing a variety of sleep recordings is treated. The technique is demonstrated on seven all-night recordings resulting in 89.6% accuracy for about 5000 pages of data.

Effects of methionine hydroxy analog on milk secretion and ruminal and blood variables of dairy cows fed a low fiber diet.

Twenty-eight Holstein cows were in a reversal trial to determine effects of adding 34 g/day of methionine hydroxy analog to a low fiber ration. Cows were subjected to this feeding treatment sequence after parturition: days 1 through 28, normal fiber ration; days 29 through 56, low fiber ration; days 57 through 77, low fiber ration with or without 34 g/day of methionine hydroxy analog; days 78 through 98, low fiber ration; days 99 through 119, low fiber ration with or without 34 g/day of methionine hydroxy analog (days 57 through 77 treatments reversed). Milk, rumen fluid, and blood were collected at regular intervals during each period. Overall, supplementation with methionine hydroxy analog increased only milk fat 6%. After milk fat percentage was depressed by low fiber diet, supplementation with methionine hydroxy analog increased milk fat percentage only for those cows that initially had less than 20% decrease of milk fat percentage. This enhanced milk fat percentage was accompanied by increase of ratio of ruminal acetate to propionate in animals with less than 10% depression. Supplementation with methionine hydroxy analog had no effect on blood glucose, insulin, or acetate concentrations and did not affect uptake of acetate by the mammary gland. Initial milk fat production affects the ability of methionine hydroxy analog supplementation to enhance milk fat percentage of dairy cattle fed low fiber diets.