Predicting evaporation rate is one of important elements for hydrology planning. There are several methods to estimate evaporation from a water surface. The objective of this study was to test the capability of artifi...Predicting evaporation rate is one of important elements for hydrology planning. There are several methods to estimate evaporation from a water surface. The objective of this study was to test the capability of artificial neural networks （ANNs） to predict evaporation using 10 years data set （1999 to 2008） from Ahvaz meteorological station and has been compared with values obtained using pan evaporation. Software Qnet 2000 has been utilized to model the evaporation. The Qnet 2000 was trained with monthly climate data （Solar radiation, minimum and maximum temperature, minimum and maximum relative humidity, and wind velocity） as input. The model was approximately implemented 144 times that finally hyperbolic secant stimulant function of 4 input parameters including minimum temperature, maximum temperature, solar radiation and wind velocity and 6 nodes in hidden layer has been yielded the best outcome. Correlation coefficients （R2） in training and testing sections are to 97.4% and 97.3% respectively. Also maximum errors in training and testing sections equaled to 18% and 24% respectively. Results showed ANNs approach works well for the data set used in this region.展开更多
The interaction of nanoparticles with proteins is extremely complex, important for understanding the biological properties of nanomaterials, but is very poorly understood. We have employed a combinatorial library of s...The interaction of nanoparticles with proteins is extremely complex, important for understanding the biological properties of nanomaterials, but is very poorly understood. We have employed a combinatorial library of surface modified gold nanoparticles to interrogate the relationships between the nanoparticle surface chemistry and the specific and nonspecific binding to a common, important, and representative enzyme, acetylcholinesterase （ACHE）. We also used Bayesian neural networks to generate robust quantitative structure-property relationship （QSPR） models relating the nanoparticle surface to the AChE binding that also provided significant understanding into the molecular basis for these interactions. The results illustrate the insights that result from a synergistic blending of experimental combinatorial synthesis and biological testing of nanoparticles with quantitative computational methods and molecular modeling.展开更多
文摘Predicting evaporation rate is one of important elements for hydrology planning. There are several methods to estimate evaporation from a water surface. The objective of this study was to test the capability of artificial neural networks （ANNs） to predict evaporation using 10 years data set （1999 to 2008） from Ahvaz meteorological station and has been compared with values obtained using pan evaporation. Software Qnet 2000 has been utilized to model the evaporation. The Qnet 2000 was trained with monthly climate data （Solar radiation, minimum and maximum temperature, minimum and maximum relative humidity, and wind velocity） as input. The model was approximately implemented 144 times that finally hyperbolic secant stimulant function of 4 input parameters including minimum temperature, maximum temperature, solar radiation and wind velocity and 6 nodes in hidden layer has been yielded the best outcome. Correlation coefficients （R2） in training and testing sections are to 97.4% and 97.3% respectively. Also maximum errors in training and testing sections equaled to 18% and 24% respectively. Results showed ANNs approach works well for the data set used in this region.
文摘The interaction of nanoparticles with proteins is extremely complex, important for understanding the biological properties of nanomaterials, but is very poorly understood. We have employed a combinatorial library of surface modified gold nanoparticles to interrogate the relationships between the nanoparticle surface chemistry and the specific and nonspecific binding to a common, important, and representative enzyme, acetylcholinesterase （ACHE）. We also used Bayesian neural networks to generate robust quantitative structure-property relationship （QSPR） models relating the nanoparticle surface to the AChE binding that also provided significant understanding into the molecular basis for these interactions. The results illustrate the insights that result from a synergistic blending of experimental combinatorial synthesis and biological testing of nanoparticles with quantitative computational methods and molecular modeling.
