A novel dynamic batch selective sampling algorithm based on version space analysis is presented. In the traditional batch selective sampling, example selection is entirely determined by the existing unreliable classif...A novel dynamic batch selective sampling algorithm based on version space analysis is presented. In the traditional batch selective sampling, example selection is entirely determined by the existing unreliable classification boundary; meanwhile, within a batch, examples labeled previously fail to provide instructive information for the selection of the rest. As a result, using the examples selected in batch mode for model refinement will jeopardize the classification performance. Based on the duality between feature space and parameter space under the SVM active learning fi：amework, dynamic batch selective sampling is proposed to address the problem. We select a batch of examples dynamically, using the examples labeled previously as guidance for further selection. In this way, the selection of feedback examples is determined by both the existing classification model and the examples labeled previously. Encouraging experimental results demonstrate the effectiveness of the proposed algorithm.展开更多
Aiming at the topic of electroencephalogram （EEG） pattern recognition in brain computer interface （BCI）, a classification method based on probabilistic neural network （PNN） with supervised learning is presented ...Aiming at the topic of electroencephalogram （EEG） pattern recognition in brain computer interface （BCI）, a classification method based on probabilistic neural network （PNN） with supervised learning is presented in this paper. It applies the recognition rate of training samples to the learning progress of network parameters. The learning vector quantization is employed to group training samples and the Genetic algorithm （GA） is used for training the network＇ s smoothing parameters and hidden central vector for detemlining hidden neurons. Utilizing the standard dataset I （a） of BCI Competition 2003 and comparing with other classification methods, the experiment results show that the best performance of pattern recognition Js got in this way, and the classification accuracy can reach to 93.8%, which improves over 5% compared with the best result （88.7 % ） of the competition. This technology provides an effective way to EEG classification in practical system of BCI.展开更多
The integrated effect of irrigation and agricultural practices on soil salinity in the Jordan Valley （JV）, where over 60% of Jordan＇s agricultural produce is grown, was investigated in this study during 2009-2010. ...The integrated effect of irrigation and agricultural practices on soil salinity in the Jordan Valley （JV）, where over 60% of Jordan＇s agricultural produce is grown, was investigated in this study during 2009-2010. Due to the differences in agricultural operations, cropping patterns, irrigation management, and weather conditions, 206 top- and sub-soil samples were taken every 1 to 3 km from representative farms along a north-south （N-S） transect with 1 to 2 km lateral extents. Soil electrical conductivity of saturated extract （ECse）, Ca, Mg, K, Na, CI, and Na adsorption ratio （SAR） were determined in saturated paste extracts. Results indicated that about 63% of soils in the JV are indeed saline, out of which almost 46% are moderately to strongly saline. Along the N-S transect of the JV, ECse increased from 4.5 to 14.1 dS m-1 in top-soil samples. Similar increase was observed for the sub-soil samples. The major chemical components of soil salinity; i.e., Ca, Mg, and C1, also showed a similar increase along the N-S transect of the valley. Moreover, compared to previous field sampling, results showed that changes in soil salinity in the JV were dramatic. In addition, it was found that C1 imposed an existing and potential threat to sensitive crops in 60% of the soils in the JV, where C1 concentrations were greater than 710 mg L-1. Under the prevalent arid Mediterranean conditions, improving the management of .irrigation water, crops, and nutrient inputs and increasing water and fertilizer use efficiencies should be indispensable to conserve and sustain the already fragile agricultural soils in the JV.展开更多
In order to investigate sample minimization for classification of supercritical and subcritical patterns in supersonic inlet, three optimization methods, namely, opposite one towards nearest method, closest one toward...In order to investigate sample minimization for classification of supercritical and subcritical patterns in supersonic inlet, three optimization methods, namely, opposite one towards nearest method, closest one towards the byper-plane method and random selection method, are proposed for investigation on minimization of classification samples for supercritical and subcritical patterns of supersonic inlet. The study has been carried out to analyze wind tunnel test data and to compare the classification accuracy based on those three methods with or without priori knowledge. Those three methods are different from each other by different selecting methods for samples. The results show that one of the optimization methods needs the minimization samples to get the highest classification accuracy without priori knowledge. Meanwhile, the number of minimization samples needed to get highest classification accuracy can be further reduced by introducing priori knowledge. Furthermore, it demonstrates that the best optimization method has been found by comparing all cases studied with or without introducing priori knowledge. This method can be applied to reduce the number of wind tunnel tests to obtain the inlet performance and to identify the supercritical/subcritical modes for supersonic inlet.展开更多
文摘A novel dynamic batch selective sampling algorithm based on version space analysis is presented. In the traditional batch selective sampling, example selection is entirely determined by the existing unreliable classification boundary; meanwhile, within a batch, examples labeled previously fail to provide instructive information for the selection of the rest. As a result, using the examples selected in batch mode for model refinement will jeopardize the classification performance. Based on the duality between feature space and parameter space under the SVM active learning fi：amework, dynamic batch selective sampling is proposed to address the problem. We select a batch of examples dynamically, using the examples labeled previously as guidance for further selection. In this way, the selection of feedback examples is determined by both the existing classification model and the examples labeled previously. Encouraging experimental results demonstrate the effectiveness of the proposed algorithm.
基金Supported by the National Natural Science Foundation of China (No. 30570485)the Shanghai "Chen Guang" Project (No. 09CG69).
文摘Aiming at the topic of electroencephalogram （EEG） pattern recognition in brain computer interface （BCI）, a classification method based on probabilistic neural network （PNN） with supervised learning is presented in this paper. It applies the recognition rate of training samples to the learning progress of network parameters. The learning vector quantization is employed to group training samples and the Genetic algorithm （GA） is used for training the network＇ s smoothing parameters and hidden central vector for detemlining hidden neurons. Utilizing the standard dataset I （a） of BCI Competition 2003 and comparing with other classification methods, the experiment results show that the best performance of pattern recognition Js got in this way, and the classification accuracy can reach to 93.8%, which improves over 5% compared with the best result （88.7 % ） of the competition. This technology provides an effective way to EEG classification in practical system of BCI.
文摘选取两组有机化合物的熔点数据,采用ADMEWORKS Model Buider软件计算并选取描述符,以所选描述符为自变量,熔点为因变量。通过三种模式识别方法——K-最近邻法（K-Nearest Neighbor Method,KNN）、K-均值聚类法（K-Means Clustering Method,KMC）和投影寻踪方法（Projection Pursuit Method,PP）对样本进行分类,将分类后的样本分别以多元线性回归（Multiple Linear Regression,MLR）、偏最小二乘（Partial Least Squares,PLS）和人工神经网络（Artificial Neural Network,ANN）建立QSPR模型。结果表明,三种模式识别方法均可以提高模型的预测能力。模型的预测能力不仅与结构相似度有关,还与建模方法有关。非线性模型预测能力要优于线性模型。
文摘The integrated effect of irrigation and agricultural practices on soil salinity in the Jordan Valley （JV）, where over 60% of Jordan＇s agricultural produce is grown, was investigated in this study during 2009-2010. Due to the differences in agricultural operations, cropping patterns, irrigation management, and weather conditions, 206 top- and sub-soil samples were taken every 1 to 3 km from representative farms along a north-south （N-S） transect with 1 to 2 km lateral extents. Soil electrical conductivity of saturated extract （ECse）, Ca, Mg, K, Na, CI, and Na adsorption ratio （SAR） were determined in saturated paste extracts. Results indicated that about 63% of soils in the JV are indeed saline, out of which almost 46% are moderately to strongly saline. Along the N-S transect of the JV, ECse increased from 4.5 to 14.1 dS m-1 in top-soil samples. Similar increase was observed for the sub-soil samples. The major chemical components of soil salinity; i.e., Ca, Mg, and C1, also showed a similar increase along the N-S transect of the valley. Moreover, compared to previous field sampling, results showed that changes in soil salinity in the JV were dramatic. In addition, it was found that C1 imposed an existing and potential threat to sensitive crops in 60% of the soils in the JV, where C1 concentrations were greater than 710 mg L-1. Under the prevalent arid Mediterranean conditions, improving the management of .irrigation water, crops, and nutrient inputs and increasing water and fertilizer use efficiencies should be indispensable to conserve and sustain the already fragile agricultural soils in the JV.
基金Academy of Fundamental and Interdisciplinary Sciences,Harbin Institute of Technology
文摘In order to investigate sample minimization for classification of supercritical and subcritical patterns in supersonic inlet, three optimization methods, namely, opposite one towards nearest method, closest one towards the byper-plane method and random selection method, are proposed for investigation on minimization of classification samples for supercritical and subcritical patterns of supersonic inlet. The study has been carried out to analyze wind tunnel test data and to compare the classification accuracy based on those three methods with or without priori knowledge. Those three methods are different from each other by different selecting methods for samples. The results show that one of the optimization methods needs the minimization samples to get the highest classification accuracy without priori knowledge. Meanwhile, the number of minimization samples needed to get highest classification accuracy can be further reduced by introducing priori knowledge. Furthermore, it demonstrates that the best optimization method has been found by comparing all cases studied with or without introducing priori knowledge. This method can be applied to reduce the number of wind tunnel tests to obtain the inlet performance and to identify the supercritical/subcritical modes for supersonic inlet.
