This research presents bending responses of FG-GPLRC plates based upon higher order shear deformation theory (HSDT) for various sets of boundary conditions. The rule of the mixture and modified Halpin–Tsai model are engaged to provide the effective material constant of the composite layers. By employing Hamilton’s principle, the governing equations of the structure are derived and solved with the aid of the differential quadrature method (DQM). Afterward, a parametric study is done to present the effects of three kinds of FG patterns, weight fraction of the GPLs, radius ratio, and thickness to inner radius ratio on the bending characteristics of the FG-GPLRC disk. Numerical results reveal that in the initial value of the \(Zt/h\), using more GPLs for reinforcing the structure provides an increase in the normal stresses but this matter is inverse for the higher value of the \(Zt/h\). The results show that considering the smaller radius ratio is a reason for boosting the shear stresses of the structure for each \(Zt/h\). Another consequence is that for the negative value of \(Zt/h\), it is true that by increasing \(h/{R}_{i}\) , the normal stresses increases but if there is positive value for \(Zt/h\), the radial and circumferential stresses fall down by having an increase in the \(h/{R}_{i}\).
针对基于规则和统计的传统中文简历解析方法效率低、成本高、泛化能力差的缺点,提出一种基于特征融合的中文简历解析方法,即级联Word2Vec生成的词向量和用BLSTM(Bidirectional Long Short-Term Memory)建模字序列生成的词向量,然后再结合BLSTM和CRF(Conditional Random Fields)对中文简历进行解析(BLSTM-CRF)。为了提高中文简历解析的效率,级联包含字序列信息的词向量和用Word2Vec生成的词向量,融合成一个新的词向量表示;再由BLSTM强大的学习能力融合词的上下文信息,输出所有可能标签序列的分值给CRF层;再由CRF引入标签之间约束关系求解最优序列。利用梯度下降算法训练神经网络,使用预先训练的词向量和Dropout优化神经网络,最终完成对中文简历的解析工作。实验结果表明,所提的特征融合方法优于传统的简历解析方法。 相似文献