Method for improving classification performance of neural network based on fuzzy input and network inversion

In order to effectively improve the classification performance of neural network, first architecture of fuzzy neural network with fuzzy input was proposed. Next a cost function of fuzzy outputs and non-fuzzy targets was defined. Then a learning algorithm from the cost function for adjusting weights was derived. And then the fuzzy neural network was inversed and fuzzified inversion algorithm was proposed. Finally, computer simulations on real-world pattern classification problems examine the effectives of the proposed approach. The experiment results show that the proposed approach has the merits of high learning efficiency, high classification accuracy and high generalization capability.

THE DEPENDENCE OF THE INVERSION LAYER THICKNESS ON THE FILM THICKNESS IN THIN-FILM SOI STRUCTURES

The dependence of the inversion-layer thickness on the film thickness in thin-film SOI structure is analyzed theoretically by using computer simulation. A new concept and parameter, the critical thickness of thin film all-bulk inversion, is introduced for the design of thin-film MOS/SOI devices. It is necessary to select the film thickness T(s1) close to the all-bulk strong inversion critical thickness in order to get high-speed and high-power operation of ultra-thin film MOS/SOI devices.

Surface transformation and inversion domain boundaries in gallium nitride nanorods

Phase transformation and subdomain structure in [0001]-oriented gallium nitride (GaN) nanorods of different sizes are studied using molecular dynamics simulations. The analysis concerns the structure of GaN nanorods at 300 K without external loading. Calculations show that a transformation from wurtzite to a tetragonal structure occurs along {0110} lateral surfaces, leading to the formation of a six-sided columnar inversion domain boundary (IDB) in the [0001] direction of the nanorods. This structural configuration is similar to the IDB structure observed experimentally in GaN epitaxial layers. The transformation is significantly dependent on the size of the nanorods.

Investigations of phase inversion and frictional pressure gradients in upward and downward oil–water flow in vertical pipes

The present study has attempted to investigate phase inversion and frictional pressure gradients during simultaneous vertical flow of oil and water two-phase through upward and downward pipes. The liquids selected were white oil (44 mPa s viscosity and 860 kg/m3 density) and water. The measurements were made for phase velocities varying from 0 to 1.24 m/s for water and from 0 to 1.87 m/s for oil, respectively. Experiments were carried either by keeping the mixture velocity constant and increasing the dispersed phase fraction or by keeping the continuous phase superficial velocity constant and increasing the dispersed phase superficial velocity. From the experimental results, it is shown that the frictional pressure gradient reaches to its lower value at the phase inversion point in this work. The points of phase inversion are always close to an input oil fraction of 0.8 for upward flow and of 0.75 for downward flow, respectively. A few models published in the literature are used to predict the phase inversion point and to compare the results with available experimental data. Suitable methods are suggested to predict the critical oil holdup at phase inversion based on the different viscosity ratio ranges. Furthermore, the frictional pressure gradient is analyzed with several suitable theoretical models according to the existing flow patterns. The analysis reveals that both the theoretical curves and the experimental data exhibit the same trend and the overall agreement of predicted values with experimental data is good, especially for a high oil fraction.