: This thesis applies neural network feature selection techniques to multivariate time series data to improve prediction of a target time series. Two approaches to feature selection are used. First, a subset enumeration method is used to determine which financial indicators are most useful for aiding in prediction of the S&P 500 futures daily price. The candidate indicators evaluated include RSI, Stochastics and several moving averages. Results indicate that the Stochastics and RSI indicators result in better prediction results than the moving averages. The second approach to feature selection is calculation of individual saliency metrics. A new decision boundary-based individual saliency metric, and a classifier independent saliency metric are developed and tested. Ruck's saliency metric, the decision boundary based saliency metric, and the classifier independent saliency metric are compared for a data set consisting of the RSI and Stochastics indicators as well as delayed closing price values. The decision based metric and the Ruck metric results are similar, but the classifier independent metric agrees with neither of the other metrics. The nine most salient features, determined by the decision boundary based metric, are used to train a neural network and the results are presented and compared to other published results. (AN)

/pdf/nonlinear-time-series-analysis-1j3a35n4y3.pdf

Nonlinear Time Series Analysis.

Nonlinearity is important and ubiquitous in ecology. Though detectable in principle, nonlinear behavior is often difficult to characterize, analyze, and incorporate mechanistically into models of ecosystem function. One obvious reason is that quantitative nonlinear analysis tools are data intensive (require long time series), and time series in ecology are generally short. Here we demonstrate a useful method that circumvents data limitation and reduces sampling error by combining ecologically similar multispecies time series into one long time series. With this technique, individual ecological time series containing as few as 20 data points can be mined for such important information as (1) significantly improved forecast ability, (2) the presence and location of nonlinearity, and (3) the effective dimensionality (the number of relevant variables) of an ecological system.

/pdf/extending-nonlinear-analysis-to-short-ecological-time-series-1ut6lnvvip.pdf

Extending nonlinear analysis to short ecological time series.

The spatial and temporal structure of predator-prey relationships in the Celtic Sea was investigated for 4 commercially important predator species (cod Gadus morhua, hake Merluccius merluccius, megrim Lepidorhombus whiffiagonis, and whiting Merlangius merlangus) using stomach- content and bottom-trawl survey data for the period 1982 to 1995. Blue whiting Micromesistius poutassou were consumed more often during the summer months, whereas mackerel Scomber scombrus and Trisopterus spp. (poor cod T. minutus, Norway pout T. esmarkii, and bib T. luscus) were found more often in predator stomachs during the winter half-year. On a spatial scale, blue whiting was consumed over the shelf edge, in accordance with their higher densities in the environ- ment, while mackerel, horse mackerel Trachurus trachurus and Trisopterus spp. were eaten more often on the continental shelf, again in agreement with their depth-related density-distribution patterns. The spatial distribution of whiting closely matched that of a key prey, Trisopterus spp. This might indicate an 'aggregative response' by the predator. The results of this study suggest spatial and seasonal prey-switching behaviour by cod, hake and whiting. Overall, the Celtic Sea fish community is characterised by opportunistic predators and general independence between predator and prey distributions. Interspecific predator interactions are reduced by size-, space- and time-dependent feeding behaviours.

/pdf/spatial-and-temporal-structure-of-predator-prey-f4yop2hfw6.pdf

Spatial and temporal structure of predator–prey relationships in the Celtic Sea fish community

Pelagic North Sea whiting Merlangius merlangus preyed upon (pelagic) sprat and sandeel, while demersal whiting preyed upon (demersal) Norway pout and herring. Values of diet breadth were low for both feeding groups using Levin's index. Diet overlaps within layers were low (D 0·74) using Schoener's index. Selection of prey was density dependent. However, prey size also played an important role. The diet of whiting shifted from amphipods and mysids to fish with increasing predator length, and the length of prey consumed increased significantly with length of whiting. The fact that the stomach contents differed between the feeding groups demonstrates the need for a sampling design that includes both pelagic and demersal habitats when trying to quantify the diet of whiting.

Diet comparison between pelagic and demersal whiting in the North Sea

This study presents an analysis of the bias introduced by using simplified methods to calculate food intake of fish from stomach contents. Three sources of bias were considered: (1) the effect of e...

/pdf/bias-in-estimating-food-consumption-of-fish-by-stomach-4i7x6yiion.pdf

Bias in estimating food consumption of fish by stomach-content analysis

Food consumption of whiting, Merlangius merlangus, in the Eastern Irish Sea

Feeding behaviour of whiting, Merlangius merlangus, L. in captivity

Nonlinear time series analysis of food intake in the dab and the rainbow trout

In this study, gastric emptying and factors affecting gastric emptying in fish are briefly reviewed. How stomach-sampling time affects gastric emptying curves in relation to realising the chyme in fish is presented. A new modelling approach describing gastric emptying in fishes is also given.

https://journals.tubitak.gov.tr/veterinary/issues/vet-03-27-5/vet-27-5-1-0111-39.pdf

D.J. Grove

Papers

Food consumption of whiting, Merlangius merlangus, in the Eastern Irish Sea

Feeding behaviour of whiting, Merlangius merlangus, L. in captivity

Nonlinear time series analysis of food intake in the dab and the rainbow trout

A New Approach in Modelling Gastric Emptying in Fish