463 research outputs found
Sparse prototype representation by core sets
No full textSchleif F-M, Zhu X, Hammer B. Sparse prototype representation by core sets. In: Hujun Yin et.al, ed. IDEAL 2013. 2013
Decoding population neuronal responses by topological clustering
No full textIn this paper the use of topological clustering for decoding population neuronal responses and reducing stimulus features is described. The discrete spike trains, recorded in rat somatosensory cortex in response to sinusoidal vibrissal stimulations characterised by different frequencies and amplitudes, are first interpreted to continuous temporal activities by convolving with a decaying exponential filter. Then the self-organising map is utilised to cluster the continuous responses. The result is a topologically ordered clustering of the responses with respect to the stimuli. The clustering is formed mainly along the product of amplitude and frequency of the stimuli. Such grouping agrees with the energy coding result obtained previously based on spike counts and mutual information. To further investigate how the clustering preserves information, the mutual information between resulting stimulus grouping and responses has been calculated. The cumulative mutual information of the clustering resembles closely that of the energy grouping. It suggests that topological clustering can naturally find underlying stimulus-response patterns and preserve information among the clusters. © 2008 Springer-Verlag Berlin Heidelberg
Multivoxel pattern analysis using information-preserving EMD
No full textThis paper presents a quantitative analysis on fMRI data using the information-preserving mode decomposition. Multivoxel patterns in fMRI responses in a cognitive experiment were analyzed for spatial selectivity to color perceptions of neurons in the Lateral Geniculate Nucleus (LGN) and the primary visual cortex (V1). The performance of the new method is tested and evaluated in a case study and the results are compared with the previous findings on the same dataset. While conforming to the previous study, the new results have shown improved classification of patterns for unique hues in V1
Generalized derivative based Kernelized learning vector quantization
Full text linkSchleif F-M, Villmann T, Hammer B, Schneider P, Biehl M. Generalized derivative based Kernelized learning vector quantization. In: Fyfe C, Tino P, Charles D, Garcia-Osorio C, Yin H, eds. Intelligent Data Engineering and Automated Learning – IDEAL 2010 11th International Conference, Paisley, UK, September 1-3, 2010. Proceedings. Berlin u.a.: Springer; 2010: 21-28
Towards Spectral-Texture Approach to Hyperspectral Image Analysis for Plant Classification
No full textThe use of hyperspectral imaging systems in studying plant properties, types, and conditions has significantly increased due to numerous economical and financial benefits. It can also enable automatic identification of plant phenotypes. Such systems can underpin a new generation of precision agriculture techniques, for instance, the selective application of plant nutrients to crops, preventing costly losses to soils, and the associated environmental impact to their ingress into watercourses. This paper is concerned with the analysis of hyperspectral images and data for monitoring and classifying plant conditions. A spectral-texture approach based on feature selection and the Markov random field model is proposed to enhance classification and prediction performance, as compared to conventional approaches. Two independent hyperspectral datasets, captured by two proximal hyperspectral instrumentations with different acquisition dates and exposure times, were used in the evaluation. Experimental results show promising improvements in the discrimination performance of the proposed approach. The study shows that such an approach can shed a light on the attributes that can better differentiate plants, their properties, and conditions
Towards Spectral-Texture Approach to Hyperspectral Image Analysis for Plant Classification
Full text linkThe use of hyperspectral imaging systems in studying plant properties, types, and conditions has significantly increased due to numerous economical and financial benefits. It can also enable automatic identification of plant phenotypes. Such systems can underpin a new generation of precision agriculture techniques, for instance, the selective application of plant nutrients to crops, preventing costly losses to soils, and the associated environmental impact to their ingress into watercourses. This paper is concerned with the analysis of hyperspectral images and data for monitoring and classifying plant conditions. A spectral-texture approach based on feature selection and the Markov random field model is proposed to enhance classification and prediction performance, as compared to conventional approaches. Two independent hyperspectral datasets, captured by two proximal hyperspectral instrumentations with different acquisition dates and exposure times, were used in the evaluation. Experimental results show promising improvements in the discrimination performance of the proposed approach. The study shows that such an approach can shed a light on the attributes that can better differentiate plants, their properties, and conditions
Towards Spectral-Texture Approach to Hyperspectral Image Analysis for Plant Classification
Full text linkThe use of hyperspectral imaging systems in studying plant properties, types, and conditions has significantly increased due to numerous economical and financial benefits. It can also enable automatic identification of plant phenotypes. Such systems can underpin a new generation of precision agriculture techniques, for instance, the selective application of plant nutrients to crops, preventing costly losses to soils, and the associated environmental impact to their ingress into watercourses. This paper is concerned with the analysis of hyperspectral images and data for monitoring and classifying plant conditions. A spectral-texture approach based on feature selection and the Markov random field model is proposed to enhance classification and prediction performance, as compared to conventional approaches. Two independent hyperspectral datasets, captured by two proximal hyperspectral instrumentations with different acquisition dates and exposure times, were used in the evaluation. Experimental results show promising improvements in the discrimination performance of the proposed approach. The study shows that such an approach can shed a light on the attributes that can better differentiate plants, their properties, and conditions
Towards Spectral-Texture Approach to Hyperspectral Image Analysis for Plant Classification
Full text linkThe use of hyperspectral imaging systems in studying plant properties, types, and conditions has significantly increased due to numerous economical and financial benefits. It can also enable automatic identification of plant phenotypes. Such systems can underpin a new generation of precision agriculture techniques, for instance, the selective application of plant nutrients to crops, preventing costly losses to soils, and the associated environmental impact to their ingress into watercourses. This paper is concerned with the analysis of hyperspectral images and data for monitoring and classifying plant conditions. A spectral-texture approach based on feature selection and the Markov random field model is proposed to enhance classification and prediction performance, as compared to conventional approaches. Two independent hyperspectral datasets, captured by two proximal hyperspectral instrumentations with different acquisition dates and exposure times, were used in the evaluation. Experimental results show promising improvements in the discrimination performance of the proposed approach. The study shows that such an approach can shed a light on the attributes that can better differentiate plants, their properties, and conditions
Modeling gene expression time-series with radial basis function neural networks
No full textGene expression time-series are discrete, noisy, short and usually unevenly sample& Most existing methods used to "npare.expression -profiles operate directly on the time points. While.modelling the profiles can lead to more generalised, smooth characterisation of gene expressions. In this paper.a Radial Basis Function neural network is employed to model .gene expression time-series. The .Orthogonal Least Square method, used for
selection of centres, is further combined with a width optimisation scheme. The experiments on a number of expression dalasets have shown the advantages of the approach in terms of generalisation and approximation. The results on ,known datasets have indeed.coincided with biological interpretations
On Random-Forest-Based Prediction Intervals
No full textIn the context of predicting continuous variables, many proposals in the literature exist dealing with point predictions. However, these predictions have inherent errors which should be quantified. Prediction intervals (PI) are a great alternative to point predictions, as they permit measuring the uncertainty of the prediction. In this paper, we review Quantile Regression Forests and propose five new alternatives based on them, as well as on classical random forests and linear and quantile regression, for the computation of PIs. Moreover, we perform several numerical experiments to evaluate the performance of the reviewed and proposed methods and extract some guidelines on the method to choose depending on the size of the data set and the shape of the target variable.Depto. de Estadística y Ciencia de los DatosFac. de Estudios EstadísticosTRUEpu
- …
