Search results

When dealing with high dimensional data, useful information can be overwhelmed by large quantities of redundant data and relevant redundant information may take up much storage space and may be time consuming.
Combining dimension reduction of the high dimensional data with neural network organically is one of the hot spots in neural network algorithm improvement, which has made remarkable achievements.
The theory of data dimension reduction methods is many, here using the factor analysis (FA) method, which regard as the promotion of principal Component Analysis(PCA).
Case Analysis This paper uses the radar ionosphere data to simulate, which is in the standard UCI data sets[12].
Dimensions of raw data are reduced by FA, which not only reduces data redundance but also eliminates influence of the relative factors.

The prediction results are compared with experiment data obtained from actual devices, we found that the different value of real experiment data and estimation data from the GM(1,1) is small and a good agreement has been obtained.
At the first, make the original experimental Vth data acted as input, afterward the grey prediction model we can get the output.
A considerable well agreement has been obtained, see Figs 5 and 6, as comparing the prediction results and the experiment data.
The measured Vth data versus channel length.
The measured Vth data versus channel width.

In addition to basic characterization, aluminium thus processed is studied in terms of fracture toughness, in order to obtain data relevant to crashworthiness that can be readily compared with alternative materials.
These advantages have resulted in Aluminium sheet already finding its way into weight reduction for automotive applications.
Weight reduction in structural or crash resistant applications requires the use of high-strength aluminium alloys: 6000 and 7000 series can reach strength/density ratios more favorable than that of steel, with a weight reduction potential of up to 20-30 % [2].
The specific essential work of fracture was determined from extrapolation of wf vs l0 data to zero ligament length (Fig. 2 right).
The good linear fitting of the wf vs l0 data (R2=0.93) confirms the reliability of the measured we and validates the applicability of the EWF method to assess the plane stress fracture toughness of thin 6061-T6 aluminium sheets.

Predicted results are compared with available experimental data.
There is an interesting feature in stiffness degradation approach that only limited amount of data is needed for obtaining reasonable results [11]. [12] reports that the reduction of bending strength of foam cored sandwich specimen is caused by the stiffness reduction of foam due to ageing of polyurethane foam during fatigue cycles. [13] investigate the static and flexural fatigue characteristics of foam core polymer composite sandwich beams.
Predicted results are compared to the available experimental data.
If such a relationship could be established, extrapolation of data at various stress levels would be derived, thus the amount of test data needed to characterize the materials behaviour [16].
Fatigue data were generated at load levels of 100, 90, 80, 70 and 60% of the static ultimate load.

To reduce the complexity of caculation, it is necessary to extract the feature of sample data before the modeling of neural network, that is, in terms of a certain evaluation criteria, the best feature subset is selected from input feature set.
Suppose the independent variables are , the dependent variable is , and take sample points to construct the data table and .
Then can represent data table as good as possible, and has the best explanation ability to the dependent variable .
Table 1 show the partial information of selected data sets.
Table 1 Partial information of selected data set Table 2 Modleing ability comparison of three data set Conclusion PLS can reduce the input dimensions, noise pollution and multiple correlations between independent variables, namely it has a better feature extraction ability.

The real SAR image data is used to test the presented method and the experimental results verify that MABF is feasible and effective.
The simplest approaches to speckle reduction are based on temporal averaging [2].median filtering.
Then many adaptive filters for speckle reduction have been developed.
For remote sensing image data is very large, so the efficiency of different filtering algorithms is very important indicator.
It is evident from the table 1 and table 2 that MABF method is more successful in the noise reduction than other filters.

The experimental data from Cr and Cu were fitted to adsorption isotherms models.
The Cr adsorption data with various adsorbents fitted by the Langmuir and Freundlich models is shown in Fig. 4.
Fig.4 Compare different experimental data with isotherm models for Cr adsorption using a) nZVI, b) nZVI- PAC, and c) nZVI- GAC The principal mechanism of Cr removal by nZVI is the oxidation reaction of iron and the reduction of Cr, resulting in the deposition of trivalent Cr and other products.
As described previously, the data obtained from Cr adsorption by nZVI were not even fitted to the Langmuir model.
After fitting the experimental data, the results were similar to the current study.

Conversely, the addition of Mn causes a slight reduction in the size and volume fraction of the CCP.
This is evident from the data of Figures 2 and 3 which illustrate the strain to fracture in the Mn and Fe variant alloys respectively, as a function of ageing condition and therefore yield strength.
In the Fe variant alloys even the smallest Fe increase halts the reduction in fin the alloy caused by peak ageing.
Discussion The fracture results of Figures 2 to 4 can be understood by analyzing the influence that Mn and Fe additions have on the microstructure of the alloys in question (i.e. the data of Tables 3 and 4).
The reduction in grain size induced by the Mn is much less significant.

Experimental data showed that the UEWS point varies for each winding angle, and the difference becomes even more pronounced, especially when the angles deviated from the ideal ±55°.
Nonetheless, the weeping stress at final failure for pipe samples with winding angle of ±57.5° exhibits notable reduction by almost 14% from 233MPa to 201MPa.
The UEWS strength was observed to degrade at winding angles of ±42° and ±73°, yielding over 40% reduction for both cases.
The comparison of the experimental strain data and those computed through laminate theory is presented in Table 2 below and graphically in Figure 2; Table 2 : Comparison of experimental data with lamination theory   Experimental data Lamination Theory (All values calculated at experimental UEWS test results!)
Fig. 2 : Comparison between UEWS experimental data and Tsai Wu failure criterion Conclusions Effects of various winding angles in biaxial UEWS tests were investigated.

Among the benefits are lowering overall cost in changing the battery and also avoid the disruption of conveying the data during battery replacement process.
Fig. 3(a) shows the plot using Table I and Table II data as starting size for this series of simulation.
Overall percent of reduction in wn is around 80% at 5 µm ZnO from 20 µm.
This plot shows the reduction of the wn with the decrease of the ZnO thickness.
The reduction of the width in both TPM and BEM is not affecting much in term of lowering wn as shown in Fig. 3.