Earticle

Calculating semantic similarity between concepts in ontology is an important issue in natural language processing and so on, so how to measure the similarity becomes a hot topic among many scholars. However, most existing methods cannot distinguish the similarity further. Confronting this problem, we propose a new semantic similarity method combining hierarchy structure of ontology and information content of two concepts based on domain ontology, which highlights the semantic information of leaves in the ontology structure. Our experiment demonstrates that, compared with other available methods, our proposal can improve the accuracy between two leaves and between leaf and non-leaf.

Surface affected layer characteristic is an important index to check machined surface integrity, which has great influence on work piece fatigue and wear resistance. Based on the definition of four filed characteristics parameters-filed gradient, field impact depth, filed mean value and field peak value, the combination of quantitative and qualitative description of high-speed milling process physical field is realized. According to finite element analysis method, characteristic parameters of stress filed and temperature filed in the process of hardened steel high-speed milling are calculated, and the field characteristic parameters changing regulation under different cutting parameters is analyzed. Based on micro hardness experiment results of the surface layer and sub surface layer materials of machined surface obtained from experiment and the surface residual stress distribution obtained from data simulation, surface affected layer character, such as residual stress and hardness changing, forming mechanism and altering regulation is discovered based on filed characteristic parameter analysis. The results show that the residual stress along the depth direction is compression stress, the stress field peak value and the temperature field gradient have greatest influence on residual stress among these filed characteristic parameters; stress field peak value can be used as the symbol of strain rate hardening, which is the main factors affecting pieces hardening along the layer depth direction.

In this paper, various kinds of sub-word lexica are thoroughly investigated under the framework of Uyghur LVCSR system. Experimental results show that it is inefficient to directly model based on word units or small units like morpheme or even syllable units. It is observed that an optimal sub-word unit set between word and morpheme units can better fit for ASR system. In order to select best unit set we have investigated several effective unit segmentation, concatenation approaches, and their ASR performances. For segmentation approach, we investigate a supervised segmentation which split words into the smallest functional units - the linguistic morphemes, and an unsupervised segmentation which extract pseudo-morphemes (or statistical morphemes). In supervised model, a leaning algorithm is trained on a manually prepared training corpus, and morpho-phonetics changes are analyzed. In the unsupervised model, the Morfessor tool is used to extract pseudo-morphemes from a raw text corpus. For concatenation approach, several effective concatenation approaches are investigated based on linguistic morphemes. First is the data-driven approach which concatenates morpheme sequences based on certain measures like co-occurrence frequency or mutual probability. Second is a model based approach which merges units with global statistical criteria. In this study, the Morfessor program is revised and turned into concatenation program by controlling segmentation points. Third is the two-layer-lexica based concatenation approach which extracts an optimal sub-word unit set by aligning and comparing the ASR results of word and morpheme two lexical layers. This method utilizes both speech and text, and produced the best results in terms of WER and lexicon size, and proved to be very stable. The best optimal lexicon, which is obtained totally on the basis of HMM based acoustic model, outperformed all other baseline lexica. And when all these lexica are directly incorporated with a deep neural network (DNN) based acoustic model, without changing the speech and text training corpora and language models, the optimal lexicon not only drastically improved the ASR accuracy but also outperformed other units as a proof of the generality of the two-layer-lexica based approach.

Particle swarm optimization algorithm is a species of intelligent algorithm, it can solve the problem of multiple end of decision making. But the algorithm is based on each group of particles would have been the effective information hypothesis. For most of the optimization problem, by the convergence speed, set the parameters of the limit, so this paper proposes a new more volume particle group algorithm. Crowding mechanism algorithm was applied to select group of particles in the process of the optimal value, thus maintaining the dispersion, the selection of the global optimal value is more reasonable. To introduce the concept of half a feasible region, and then to avoid the traditional processing method only considers particles in area the disadvantages of the boundary value processing precision is not high. In respect of time complexity, the grouping method is adopted to choose random switching strategy, improve the ef

Artificial rejection system strains of particle swarm optimization algorithm, introduced selection mechanism and nonlinear optimization, using nonlinear mutation operator, such as copying and improve the algorithm convergence speed and population diversity, so as to put forward a kind of based on nonlinear plant choose the rejection of particle swarm optimization algorithm, has advantages of few parameters to adjust, and easy to implement. Combined with Iris classification problem, a weight optimization is applied to the artificial neural network, and the method based on the standard particle group algorithm and the simple comparison of artificial neural network training, the experimental results show that compared with other two kinds of algorithm performance is better, this algorithm has good convergence and stability.

In this paper an existing seven transistor (7T) CMOS SRAM cell stability is measured. N-curve method is used to find the stability of the cell. The stability parameters i.e. Static Voltage Noise Margin (SVNM), Static Current Noise Margin (SINM), Write Trip Voltage (WTV) and Write Trip Current (WTI) are measured by varying temperature and supply voltage. The existing cell has an inbuilt mechanism for charge sharing. This technique .for the write operation. The existing 7T SRAM cell has achieved 22.71% increase in stability as compared to reference cell, which validate the desired design approach.

The privacy protection model under Web service composition framework is studied in depth, then the privacy of customers can be protected effectively. Firstly, the basic framework of Web service composition is studied; Secondly, the managing method of service publication is analyzed; Thirdly, the privacy protection model under Web service composition framework is discussed; finally the evaluation method of algorithm for privacy protection is summarized.

In this paper, an efficient approach of combining additive expression tree model (AET) with hybrid evolutionary method is proposed to identify nonlinear systems. As linear variant of additive tree model, additive expression tree model is proposed to encode the mathematical formulations. For finding the optimal structure and parameters of systems, a hybrid evolutionary method integrating a new structure based evolutionary algorithm and cuckoo search is employed. We illustrate some experimental comparisons with neural network, neural network integrating fuzzy system and symbolic regression methods. Experimental results reveal that our model and optimization method perform better.