Search

The machine-learning algorithm is a technology that can learn from data. It can solve some problems that are difficult for humans to design and use deterministic programs. Among the available machine-learning algorithms for solving these problems, the Extreme Learning Machine (ELM) algorithm is well known because it only needs to set the number of the network hidden layer nodes and could generate a unique optimal solution. And, due to the randomness of its connection weights and thresholds, the simulated effect is random. Therefore, the selection of weights and thresholds needs to be optimized. In this paper, an enhanced ELM model is proposed to model the large-signal characteristics of GaN High Electron Mobility Transistors (HEMTs). The Sparrow Search Algorithm (SSA) is used to optimize the initial weights and thresholds of the ELM algorithm, which significantly improves the prediction ability. Moreover, the SSA algorithm has the characteristics of strong optimization ability and fast convergence speed, which greatly improves the feasibility of model training. Through comparing the training effects of the SSA-ELM model and the ELM model, it can be seen that the proposed SSA-ELM model improves the ability of the ELM model to simulate the large-signal characteristics of GaN HEMTs. [ABSTRACT FROM AUTHOR]

The Genetic Algorithm-Extreme Learning Machine (GA-ELM) neural network algorithm is proposed to model the relevant characteristics of GaN pseudomorphic high electron mobility transistor (P-HEMT) large signal. This algorithm solves the over-fitting problem of the Back Propagation (BP) neural network algorithm in the prediction data. It has the characteristics of fast calculation speed, so it can greatly save calculation processing time. It can also randomly generate the connection weights of the input layer, the hidden layer and the threshold of the hidden layer neurons, avoiding errors in parameter selection. In order to verify the superiority of the algorithm, the modeling effects of the BP neural network algorithm model, the Genetic Algorithm-Back Propagation (GA-BP) neural network algorithm model and the GA-ELM neural network algorithm model are compared in this paper. The results show that the proposed GA-ELM neural network algorithm model has the highest accuracy. [ABSTRACT FROM AUTHOR]

A new direct parameter-extraction technique applied to the small-signal equivalent circuit of heterojunction bipolar transistor (HBT) is presented. The method is based on first determining the parasitic elements and then the intrinsic elements analytically. Concerning the intrinsic parameters, all the elements are extracted using exact closed-form equations derived in terms of S-parameters by reducing the small-signal models to the simplified using peeling algorithm. The validity of the new extraction methodology is demonstrated by applying it to a 1 × 15 µm InP HBT device within the frequency range of 0.1-40 GHz. [ABSTRACT FROM AUTHOR]

Abstract  A total of 124 fecal specimens were collected from four deer farms in Zhengzhou City, China and examined for Cryptosporidium by Sheather’s sugar flotation technique. Cryptosporidim oocysts were detected in two 1-year-old sika deer, and one of the two specimens was genotyped by sequence and phylogenetic analyses of the small subunit ribosomal RNA (rRNA) (18S rRNA), 70-kDa heat shock protein (HSP70), actin, and Cryptosporidium oocyst wall protein (COWP) genes. Results obtained suggested that the Cryptosporidium studied belonged to Cryptosporidium cervine genotype, although slight sequence differences were noticed at the three loci. The similarities between this isolate and other Cryptosporidium cervine genotype isolates were 99.1–99.8%, 9.8%, 99.7%, and 100% at the 18S rRNA, HSP70, actin, and COWP loci, respectively. This study is the first report of Cryptosporidium infection in sika deer in China. [ABSTRACT FROM AUTHOR]