Search results for "Mean squared error"

showing 10 items of 145 documents

Use of Guided Regularized Random Forest for Biophysical Parameter Retrieval

2018

This paper introduces a feature selection method based on random forest -the Guided Regularized Random Forest (GRRF)- which can be used in classification and regression tasks. The method is based on the regularization of the information gain in the random forest nodes to obtain a subset of relevant and non-redundant features. The proposed method is used as a preliminary step In the process of retrieving biophysical parameters from a hyperspectral image. Preliminary experiments show that we can reduce the RMSE of the retrievals by around 7% for the Leaf Area Index and around 8% for the fraction of vegetation cover when compared to the results using random forest features.

Mean squared error22/3 OA procedurebusiness.industryComputer scienceFeature extractionHyperspectral images0211 other engineering and technologiesHyperspectral imagingPattern recognitionFeature selection02 engineering and technologyBiophysical parameter retrievalRegularization (mathematics)RegressionRandom forestFeature selection0202 electrical engineering electronic engineering information engineering020201 artificial intelligence & image processingArtificial intelligenceLeaf area indexbusinessRandom forest021101 geological & geomatics engineeringIGARSS 2018 - 2018 IEEE International Geoscience and Remote Sensing Symposium
researchProduct

Some Examples for Solving Clinical Problems Using Neural Networks

2001

In this paper neural networks are presented for solving some pharmaceutical problems. We have predicted and prevented patients with potential risk of post-Chemotherapy Emesis and potentially intoxicated patients treated with Digoxin. Neural networks have been also used for predicting Cyclosporine A concentration and Erythropoietin concentrations. Several neural networks (multilayer perceptron for classification tasks and Elman and FIR networks for prediction) and classical methods have been used. Results show how neural networks are very suitable tools for classification and prediction tasks, outperforming the classical methods. In a neural approach it is not strictly necessary to assume a …

Mean squared errorArtificial neural networkGeneralizationbusiness.industryComputer scienceMultilayer perceptronArtificial intelligencebusiness
researchProduct

Parameter optimization for amplify-and-forward relaying with imperfect channel estimation

2009

Cooperative diversity is a promising technology for future wireless networks. In this paper, we consider a cooperative communication system operating in an amplify-and-forward (AF) mode with an imperfectly-known relay fading channel. It is assumed that a pilot symbol assisted modulation (PSAM) scheme with linear minimum mean square estimator (LMMSE) is used for the channel estimation. A simple and easy-to-evaluate asymptotical upper bound (AUB) of the symbol-error-rate (SER) is derived for uncoded AF cooperative systems with quadrature amplitude modulation (QAM) constellations. Based on the AUB, we propose a criterion for the choice of parameters in the PSAM scheme, i.e., the pilot spacing …

Mean squared errorChannel state informationControl theoryVDP::Technology: 500::Information and communication technology: 550::Telecommunication: 552EstimatorFadingUpper and lower boundsAlgorithmQuadrature amplitude modulationComplex normal distributionCooperative diversityMathematics
researchProduct

A three-factor optimisation strategy for micellar liquid chromatography

2000

An interpretive optimisation methodology for micellar liquid chromatography (MLC) is shown, taking into account pH, surfactant (sodium dodecyl sulphate) and organic modifier (propanol) concentration. Two objectives are considered: to develop a highly practical straightforward three-factor optimisation for practical MLC, and, in order to avoid unecessary experiments, to link two and three-factor optimisations through a step-wise construction of the experimental design at different pH levels. The whole pH range for an ODS column (from 3 to 7) is covered. The proposed strategy was thoroughly evaluated using the chromatographic data from 81 experimental mobile phases, applied to the separation …

Mean squared errorChemistryOrganic ChemistryClinical BiochemistryAnalytical chemistryBiochemistryHigh-performance liquid chromatographyMicellar electrokinetic chromatographyAnalytical ChemistrySet (abstract data type)ChemometricsPropanolchemistry.chemical_compoundMicellar liquid chromatographyTest setBiological systemChromatographia
researchProduct

Advantages of fitting contrast curves using logistic function: a technical note.

2013

Objective The aim of this article is to demonstrate how the contrast properties of an imaging system can be ideally fitted with the use of stripe patterns and the logistic function. Study Design Stripe patterns with defined amounts of line pairs (lp/mm) per mm (10-20 lp/mm) were recorded with the use of digital photostimulable storage phosphor. Scan data and normalized image data were analyzed with the use of ImageJ and MatLab to calculate different contrast curves. Results For original scan data, the goodness of fit was 0.0000019 (sum of squared error [SSE]). The R-square was 0.9998. For normalized data the goodness of fit was 0.0007 (SSE) and the R-square 0.998. An amount of 50% contrast …

Mean squared errorComputer scienceContrast (statistics)Image processingRadiography Dental DigitalPathology and Forensic MedicineRadiographic Image EnhancementLogistic ModelsGoodness of fitStatisticsLine (geometry)Image Processing Computer-AssistedHumansRadiology Nuclear Medicine and imagingDentistry (miscellaneous)SurgeryRadiographic Image EnhancementX-Ray Intensifying ScreensOral SurgeryLogistic functionMATLABcomputerAlgorithmAlgorithmscomputer.programming_languageOral surgery, oral medicine, oral pathology and oral radiology
researchProduct

Demultiplexing Visible and Near-Infrared Information in Single-Sensor Multispectral Imaging

2016

In this paper, we study a single-sensor imaging system that uses a multispectral filter array to spectrally sample the scene. Our system captures information in both visible and near-infrared bands of the electromagnetic spectrum. Due to manufacturing limitations, the visible filters in this system also transmit the NIR radiation. Similarly, visible light is transmitted by the NIR filter, leading to inaccurate mixed spectral measurements. We present an algorithm that resolves this issue by separating NIR and visible information. Our method achieves this goal by exploiting the correlation of multispectral images in both spatial and spectral domains. Simulation results show that the mean squa…

Mean squared errorComputer sciencebusiness.industryElectromagnetic spectrum010401 analytical chemistryMultispectral imageNear-infrared spectroscopy02 engineering and technologyFilter (signal processing)01 natural sciencesSample (graphics)0104 chemical sciencesMultispectral pattern recognitionOptics0202 electrical engineering electronic engineering information engineering020201 artificial intelligence & image processingbusinessVisible spectrumRemote sensingColor and Imaging Conference
researchProduct

Therapeutic Drug Monitoring of Kidney Transplant Recipients Using Profiled Support Vector Machines

2007

This paper proposes a twofold approach for therapeutic drug monitoring (TDM) of kidney recipients using support vector machines (SVMs), for both predicting and detecting Cyclosporine A (CyA) blood concentrations. The final goal is to build useful, robust, and ultimately understandable models for individualizing the dosage of CyA. We compare SVMs with several neural network models, such as the multilayer perceptron (MLP), the Elman recurrent network, finite/infinite impulse response networks, and neural network ARMAX approaches. In addition, we present a profile-dependent SVM (PD-SVM), which incorporates a priori knowledge in both tasks. Models are compared numerically, statistically, and in…

Mean squared errorComputer sciencecomputer.software_genreBlood concentrationmedicineElectrical and Electronic EngineeringInfinite impulse responseKidney transplantationArtificial neural networkmedicine.diagnostic_testbusiness.industryPattern recognitionmedicine.diseaseComputer Science ApplicationsHuman-Computer InteractionSupport vector machineNoiseAutoregressive modelControl and Systems EngineeringTherapeutic drug monitoringMultilayer perceptronData miningArtificial intelligencebusinesscomputerSoftwareInformation SystemsIEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews)
researchProduct

TOPS-MODE approach for the prediction of blood-brain barrier permeation.

2004

The blood-brain barrier permeation has been investigated by using a topological substructural molecular design approach (TOPS-MODE). A linear regression model was developed to predict the in vivo blood-brain partitioning coefficient on a data set of 119 compounds, treated as the logarithm of the blood-brain concentration ratio. The final model explained the 70% of the variance and it was validated through the use of an external validation set (33 compounds of the 119, MAE = 0.33), a leave-one-out crossvalidation (q(2) = 0.65, S(press) = 0.43), fivefold full crossvalidation (removing 28 compounds in each cycle, MAE = 33, RMSE = 0.43) and the prediction of +/- values for an external test set …

Mean squared errorLogarithmChemistryPharmaceutical ScienceThermodynamicsPenetration (firestop)PermeationConcentration ratioModels BiologicalPartition coefficientCapillary PermeabilityBlood-Brain BarrierPredictive Value of TestsTest setLinear regressionLinear ModelsComputer SimulationJournal of pharmaceutical sciences
researchProduct

Comparison of near and mid infrared spectroscopy as green analytical tools for the determination of total polar materials in fried oils

2017

Abstract Total polar materials (TPM) are used as an indicator of the quality in the frying oil because of high values may be harmful for human health. Spanish legislation establishes the maximum level of total polar materials for frying fats and oils for human consumption around 25% (w/w). Official methods to monitor oil quality are time consuming and use a lot of chemicals; therefore it is necessary a simple and quick analytical technique to evaluate fried oils. Transmittance near-infrared (NIR) and attenuated total reflection mid-infrared (ATR-MIR) spectroscopy measurements, combined with partial least squares (PLS) regression, offer alternatives to determine TPM in fried oils with relati…

Mean squared errorMaximum levelChemistry010401 analytical chemistryAnalytical techniqueAnalytical chemistry04 agricultural and veterinary sciencesResidual040401 food science01 natural sciencesMid infrared spectroscopy0104 chemical sciencesAnalytical Chemistry0404 agricultural biotechnologyAttenuated total reflectionPartial least squares regressionPolarFood scienceSpectroscopyMicrochemical Journal
researchProduct

Performance of TES method over urban areas at a high spatial resolution scale

2013

The Temperature and Emissivity Separation (TES) algorithm is used to retrieve the LSE and LST values from hyperspectral sensors. In this work we analyse the performance of this methodology over urban areas. Three different sources of error in the processing chain of the remote sensing imagery are detected: the algorithm itself, the atmospheric correction and the 3D structure of the urban scenes. The TITAN tool is used to model all the radiative components of the signal registered by a sensor. Results show that: first, the TES algorithm used reproduces the LSE (LST) of urban materials within an RMSE of 0.017 (0.9 K). Second, 20 % of uncertainty in the water vapour content of the total atmosp…

Mean squared errorMeteorologyEmissivityRadiative transferAtmospheric correctionEnvironmental scienceHyperspectral imagingAtmospheric modelScale (map)Image resolutionRemote sensing2013 5th Workshop on Hyperspectral Image and Signal Processing: Evolution in Remote Sensing (WHISPERS)
researchProduct