Home About us Contact
|
Home About us Contact | ||
|
|
||
Mean Absolute Error (mean + absolute_error)
Selected AbstractsImproving interpolation of daily precipitation for hydrologic modelling: spatial patterns of preferred interpolatorsHYDROLOGICAL PROCESSES, Issue 23 2009Daniel Kurtzman Abstract Detailed hydrologic models require high-resolution spatial and temporal data. This study aims at improving the spatial interpolation of daily precipitation for hydrologic models. Different parameterizations of (1) inverse distance weighted (IDW) interpolation and (2) A local weighted regression (LWR) method in which elevation is the explanatory variable and distance, elevation difference and aspect difference are weighting factors, were tested at a hilly setting in the eastern Mediterranean, using 16 years of daily data. The preferred IDW interpolation was better than the preferred LWR scheme in 27 out of 31 validation gauges (VGs) according to a criteria aimed at minimizing the absolute bias and the mean absolute error (MAE) of estimations. The choice of the IDW exponent was found to be more important than the choice of whether or not to use elevation as explanatory data in most cases. The rank of preferred interpolators in a specific VG was found to be a stable local characteristic if a sufficient number of rainy days are averaged. A spatial pattern of the preferred IDW exponents was revealed. Large exponents (3) were more effective closer to the coast line whereas small exponents (1) were more effective closer to the mountain crest. This spatial variability is consistent with previous studies that showed smaller correlation distances of daily precipitation closer to the Mediterranean coast than at the hills, attributed mainly to relatively warm sea-surface temperature resulting in more cellular convection coastward. These results suggest that spatially variable, physically based parameterization of the distance weighting function can improve the spatial interpolation of daily precipitation. Copyright © 2009 John Wiley & Sons, Ltd. [source] Daily pan evaporation modelling using multi-layer perceptrons and radial basis neural networksHYDROLOGICAL PROCESSES, Issue 2 2009Özgür Ki Abstract This paper reports on investigations of the abilities of three different artificial neural network (ANN) techniques, multi-layer perceptrons (MLP), radial basis neural networks (RBNN) and generalized regression neural networks (GRNN) to estimate daily pan evaporation. Different MLP models comprising various combinations of daily climatic variables, that is, air temperature, solar radiation, wind speed, pressure and humidity were developed to evaluate the effect of each of these variables on pan evaporation. The MLP estimates are compared with those of the RBNN and GRNN techniques. The Stephens-Stewart (SS) method is also considered for the comparison. The performances of the models are evaluated using root mean square errors (RMSE), mean absolute error (MAE) and determination coefficient (R2) statistics. Based on the comparisons, it was found that the MLP and RBNN computing techniques could be employed successfully to model the evaporation process using the available climatic data. The GRNN was found to perform better than the SS method. Copyright © 2008 John Wiley & Sons, Ltd. [source] A new recursive neural network algorithm to forecast electricity price for PJM day-ahead marketINTERNATIONAL JOURNAL OF ENERGY RESEARCH, Issue 6 2010Paras Mandal Abstract This paper evaluates the usefulness of publicly available electricity market information in predicting the hourly prices in the PJM day-ahead electricity market using recursive neural network (RNN) technique, which is based on similar days (SD) approach. RNN is a multi-step approach based on one output node, which uses the previous prediction as input for the subsequent forecasts. Comparison of forecasting performance of the proposed RNN model is done with respect to SD method and other literatures. To evaluate the accuracy of the proposed RNN approach in forecasting short-term electricity prices, different criteria are used. Mean absolute percentage error, mean absolute error and forecast mean square error (FMSE) of reasonably small values were obtained for the PJM data, which has correlation coefficient of determination (R2) of 0.7758 between load and electricity price. Error variance, one of the important performance criteria, is also calculated in order to measure robustness of the proposed RNN model. The numerical results obtained through the simulation to forecast next 24 and 72,h electricity prices show that the forecasts generated by the proposed RNN model are significantly accurate and efficient, which confirm that the proposed algorithm performs well for short-term price forecasting. Copyright © 2009 John Wiley & Sons, Ltd. [source] Sensitivity analysis of neural network parameters to improve the performance of electricity price forecastingINTERNATIONAL JOURNAL OF ENERGY RESEARCH, Issue 1 2009Paras Mandal Abstract This paper presents a sensitivity analysis of neural network (NN) parameters to improve the performance of electricity price forecasting. The presented work is an extended version of previous works done by authors to integrate NN and similar days (SD) method for predicting electricity prices. Focus here is on sensitivity analysis of NN parameters while keeping the parameters same for SD to forecast day-ahead electricity prices in the PJM market. Sensitivity analysis of NN parameters include back-propagation learning set (BP-set), learning rate (,), momentum (,) and NN learning days (dNN). The SD parameters, i.e. time framework of SD (d=45 days) and number of selected similar price days (N=5) are kept constant for all the simulated cases. Forecasting performance is carried out by choosing two different days from each season of the year 2006 and for which, the NN parameters for the base case are considered as BP-set=500, ,=0.8, ,=0.1 and dNN=45 days. Sensitivity analysis has been carried out by changing the value of BP-set (500, 1000, 1500); , (0.6, 0.8, 1.0, 1.2), , (0.1, 0.2, 0.3) and dNN (15, 30, 45 and 60 days). The most favorable value of BP-set is first found out from the sensitivity analysis followed by that of , and ,, and based on which the best value of dNN is determined. Sensitivity analysis results demonstrate that the best value of mean absolute percentage error (MAPE) is obtained when BP-set=500, ,=0.8, ,=0.1 and dNN=60 days for winter season. For spring, summer and autumn, these values are 500, 0.6, 0.1 and 45 days, respectively. MAPE, forecast mean square error and mean absolute error of reasonably small value are obtained for the PJM data, which has correlation coefficient of determination (R2) of 0.7758 between load and electricity price. Numerical results show that forecasts generated by developed NN model based on the most favorable case are accurate and efficient. Copyright © 2008 John Wiley & Sons, Ltd. [source] ORIGINAL ARTICLE: Investigation of the prediction accuracy of vancomycin concentrations determined by patient-specific parameters as estimated by Bayesian analysisJOURNAL OF CLINICAL PHARMACY & THERAPEUTICS, Issue 5 2010Y. Hiraki BSc Summary Background/Objective:, There have been many studies of therapeutic drug monitoring (TDM) of vancomycin (VCM) based on Bayesian analysis, but there have been no reports of the accuracy of prediction based on Bayesian-estimated patient-specific parameters. This study was conducted to compare the accuracy of prediction based on the population pharmacokinetic (PPK) method and Bayesian-estimated parameters. Method:, The subjects were 22 patients who were treated with VCM for MRSA infection and whose blood was sampled twice or more during the administration period. The concentrations between the blood samples were predicted based on the concentrations in the first blood samples based on the PPK method using mean parameters for the Japanese population and Bayesian-estimated patient-specific pharmacokinetic parameters. The mean prediction error (ME), mean absolute error (MAE) and root mean squared error (RMSE) were compared to examine the accuracy of prediction based on Bayesian-estimated patient-specific parameters. Results and discussion:, The mean measured VCM concentration was 10·43 ± 5·19 ,g/mL, whereas the mean concentration predicted based on the PPK method was 8·52 ± 4·34 ,g/mL, with an ME of ,1·91, MAE of 2·93 and RMSE of 3·21. The mean concentration predicted based on patient-specific parameters was 9·62 ± 4·95 ,g/mL with ME of ,0·81, MAE of 1·38 and RMSE of 1·74. The ME and MAE for the concentrations predicted using patient-specific parameters were smaller compared with those predicted using the PPK method (P = 0·0471 and 0·0003, respectively), indicating superior prediction with a significant difference between approaches. Conclusion:, Prediction using Bayesian estimates of patient-specific parameters was better than by the PPK method. However, when using patient-specific parameters it is still necessary to fully understand the clinical status of the patient and frequently determine VCM concentrations. [source] Accuracy of distributed multipoles and polarizabilities: Comparison between the LoProp and MpProp modelsJOURNAL OF COMPUTATIONAL CHEMISTRY, Issue 6 2007P. Söderhjelm Abstract Localized multipole moments up to the fifth moment as well as localized dipole polarizabilities are calculated with the MpProp and the newly developed LoProp methods for a total of 20 molecules, predominantly derived from amino acids. A comparison of electrostatic potentials calculated from the multipole expansion obtained by the two methods with ab initio results shows that both methods reproduce the electrostatic interaction with an elementary charge with a mean absolute error of ,1.5 kJ/mol at contact distance and less than 0.1 kJ/mol at distances 2 Å further out when terms up to the octupole moments are included. The polarizabilities are tested with homogenous electric fields and are found to have similar accuracy. The MpProp method gives better multipole moments unless diffuse basis sets are used, whereas LoProp gives better polarizabilities. © 2007 Wiley Periodicals, Inc. J Comput Chem, 2007 [source] | ||||||||||