The errors between output and ideal results of training samples were smaller.
样本输出结果较理想值误差较小,分类器的识别结果完全符合实际情况。
After trained network by 25 training samples, it is used to test 15 test samples.
经25个训练样本对该网络进行训练后,利用15个测试样本对网络进行了测试。
Meanwhile we divided the results into two parts: training samples and testing samples.
同时将实验结果分为两组:训练样本和测试样本。
The methods of generating training samples and alternative optimization are put forward.
提出了生成训练样本和方案优选方法。
Because of the better generalization performance of SVM, less training samples are needed.
利用支撑矢量机具有更好的推广能力,可以使用较少的训练样本。
Modeling with this method can achieve high precision if the training samples are reliable.
只要训练样本可靠,采用该方法建模可以达到比较高的精度要求。
Unfortunately, in hyper-spectral image classification, training samples are usually limited.
然而在高光谱图像分类中,训练样本通常是有限的。
The composition and training method of training samples in the neural network are presented.
阐述了该神经网络训练样本的组成和训练方法。
A periodic function, finite Fourier series, is used to activate the actuator for obtaining training samples.
用周期函数,有限项傅立叶级数,作为激励函数来获取训练样本。
In order to improve the accuracy of model prediction, the training samples should be representatively prepared.
为了提高模型的预测精度,在训练样本的选择上还应具有一定的代表性。
This paper investigates the feasibility of using analytically generated training samples to train neural networks.
本文研究了使用可分析的学习样本来训练神经网络的可行性问题。
The support vector machine is a learning algorithm, which has a good classification ability for limited training samples.
支撑矢量机是一种能在训练样本数很少的情况下达到很好分类推广能力的学习算法。
Our method gains better estimation of data distributions and mitigates the unrepresentative problem of small-size training samples.
我们的方法得到了数据分布的较好估计,能够减轻小样本的无代表性问题。
Taking the example of designing classifier in intrusion detection system, the selection of training samples for classifier is studied.
以入侵检测系统中的分类器设计为例,研究分类器训练样本选择问题。
The algorithm selects training samples by local sample density, to reduce the training samples and thus to improve the speed of learning.
该算法根据样本的局部密度选择训练样本,减少参加训练的样本数量,提高学习速度。
For the training itself, import the training samples into Workbench and make sure that the categories associated with the samples are correct.
在进行训练时,将训练样本导入Workbench中,并确保与样本相关联的目录是正确。
The correct classification rate is 100% for training samples by the two methods, and some targets are predicted as potential Sn ore-field.
二种方法对训练样本的分类正确率达100%。据此模型预报了若干个矿点为锡矿区。
With the accumulated and improved training samples, it automatically modifies the parameters of network structure and probability distribution.
该模型还可以通过不断积累完善训练样本,自动修正网络结构参数和概率分布参数。
After training samples and test samples are respectively projected towards the fusion feature space, recognition features are accordingly gained.
训练样本与测试样本分别朝融合特征空间投影,从而得到识别特征。
If the distribution of training samples only had little influence on the sub-classification, the combined classifiers would have stable performances.
当子分类器均受训练样本分布影响较小,组合结果也具有较好的稳定性。
The RBF neural networks for image reconstruction were trained in MATLAB environment. The training samples were obtained by using finite element method.
在MATLAB环境下对所研究的图像重建用RBF神经网络进行训练,并通过有限元法获得训练所需要的训练样本集。
The first order gradient optimization algorithm is employed to train the network. The training samples stem from the neutron kinetics of the point-reactor.
神经网络的训练采用一阶梯度优化算法,利用点堆中子动力学模型产生训练样本。
It depends on the stability of sub-classification that whether the results of combined classification are affected by the distribution of training samples.
组合分类结果受训练样本分布的影响取决于子分类器的稳定性。
The networks are trained by the fast BP algorithm via fuzzy variables decision, and training samples are provided by the dynamic inversion control results.
网络的训练利用改进的BP算法,将因子模糊化快速进行。样本点数据则由利用动态逆控制所得到的结果来提供。
At present there are many methods that could deal well with frontal view face recognition when there is sufficient number of representative training samples.
目前有许多处理正面视觉人脸的识别方法,当有充分数量的有代表性的样本时,能取得较好的识别效果。
This algorithm combine advantages of KNN and Clustering, decreasing training samples and quantity of algorithm calculating, and increasing the speed of retrieval.
该算法将聚类方法和KNN算法的优点结合起来,从而达到缩减了训练样本数量,减少了算法计算量,加快检索速度的目的。
From experimental results, under conditions of 200 training samples, these two characteristics maintained correct identification of a high rate at 99 and 100 percent.
从实验结果看,在200个训练样本的条件下,两种特征提取算法分别保持了99%和100%的较高识别正确率。
From experimental results, under conditions of 200 training samples, these two characteristics maintained correct identification of a high rate at 99 and 100 percent.
从实验结果看,在200个训练样本的条件下,两种特征提取算法分别保持了99%和100%的较高识别正确率。
应用推荐