Search results

Therefore, it is highly necessary in practical production to study the distribution pattern recognition of small sample censored data, especially the data estimation method for small sample censored data, to make effective use of the potential information of small sample data, and to cultivate as much useful information as possible.
The Selection of the Data Distribution Parameter.
Distribution parameter of the selected results has great impact upon the data distribution pattern recognition.
The Breakdown of the Data Distribution Pattern.
The data series were sorted, and then the series of data were censored to make the value of data not less than "2500" in the training sample set.

According to the reports[5], reduction of U(VI) to U(IV) by SRB involves at least three processes: (1) U(VI) binding to the cell surface and to extracellular biopolymers (biosorption); (2) chemical reduction of U(VI) by microbially generated hydrogen sulfide (H2S);and (3) bioreduction of U(VI), which is enzymatic dissimilatory metal reduction with U(VI) acting as a terminal electronacceptor.
As shown in Fig.3, the rate of reducing U (VI) by the reduction of electron transfer was fast, esp. in the first 6 hours.
It could be seen by comparing and analyzing the data that the U (VI) removed through the reduction mechanism of electron transfer accounts for nearly 80% of the total amount of U (VI) removed by original bacteria fluid.
SRB may detoxicate Cu2+ in some other way, which will also affect the rate of removing U (VI) by SRB. 2.4.1 Inhibition effect of Cu2+ on U (VI) reduction and removal by H2S It could be seen from Fig.6 that Cu2+ could inhibit U (VI) reduction and removal by H2S to some extent.
The reduction of U (VI) by H2S is realized by the H2S generated during the process of electron transfer.

For the variable specific surface area, using the data of Table 4 and the factorial experimental design, the Table 5 was obtained.
Table 5.24 experimental factorial designs –Estimated effects and coefficients for the data of Table 4.
Figure 3 shows the x-ray diffraction data of sample 5.
The x-ray diffraction data shows that samples 1 and 2 have 100% and 96.3% of a-alumina.
Statistics for experimenters: an introduction to design, data analysis, and model building.

However, few data are yet available to support this assumption.
Wave reduction is expected to be reliant on the density of vegetation and the surge water level.
When the waves encounter the densest vegetation, the largest rates of wave reduction occur [13][11].
Whereas a 20 years Rhizophora forest can offer up to 98 % wave reduction.
A Laboratory Study on Wave Reduction by Mangrove Forests.

The cone data and ESEM analyses results show that incorporation of a small amount of ZHS greatly promotes the char formation of PVC and decreases the amount of hazardous gases released in PVC during combustion.
Fig. 1 and Fig. 2 show that ZHS can effectively reduce the data of HRR, THR, CO and CO2 production.
The corresponding cone data are presented in Table 3.
There is a 12.9% reduction in the THR for the PVC1 sample compared to pure PVC0.
Table 3 The corresponding cone data Samples pk-HRR (kW/m2) THR (MJ/kg) pk-CO (kg/kg) pk-CO2 (kg/kg) PVC0 141.27 29.269 0.01975 0.07963 PVC1 100.43 25.504 0.00781 0.06600 * pk-HRR: The highest value of heat release rate THR: Cumulative heat energy released during flaming per unit mass of the sample.

In this report, we try to apply the SCEM to the growth rate data at various oxygen partial pressures and discuss the oxidation mechanism of SiC more in details by making a comparison between calculated growth rate curves and the observed ones.
The figures indicate that: (1) the observed data show an abrupt reduction in the thickness region less than 10 nm especially in the case of lower partial pressure, (2) after the abrupt reduction they gently reduce, (3) the calculated curves successfully reproduce the observed growth rates though restricted in the mild deceleration region, (4) the calculated curves show the rapid reduction more remarkable as higher partial pressure, which is opposite tendency to that of observed data.
The figures and point (1) indicate that the growth rate in the abrupt reduction region slightly decreases with decreasing partial pressure [5].
Although the calculated growth rate curve is still deviated from the measured data, we consider that further refined parameters will make a better fit to the observed data.
The calculation results exhibit that a rapid growth rate reduction in the initial oxidation stage eased as lower pressure, which is opposite tendency to the observed growth rates.

The benefit of using ICA is at reducing highdimensional data sets into rich accurate summary data sets.
Therefore a reduction technique such statistical data reduction techniques or independent component analysis (ICA) is required before classifier introduction.
The statistical data reduction techniques [20] used the most sensitive AE statistical measures in a windowed type fashion (every 100 pts along the STFT signal).
This data reduction technique was only applied to the AE extracted signals reference the burn and chatter phenomena trials.
GP suffers from large n-dimensional data problems and therefore requires n-dimensional reduction techniques to ensure the data applied to the classifier is both salient and rich in terms of describing the actual phenomena.

The effects of cooling rate and austenite structure on bainite formation was investigated by means of electron backscatter diffraction analysis and processing of obtained orientation data.
The data on local orientations were treated using MTEX software [8].
To choose the OR, which corresponds better to the data of the present investigation, the relationships obtained in Refs. [7, 10] were compared.
Austenite deformation results in the packet size reduction.
Note that the packet refinement does not imply a reduction of the effective grain size.

DD is a typical data-centric protocol for wireless sensor networks [1].
For instance, setting the gradient diffusion depth threshold value is used to narrow the range of the dissemination of interest messages, then to reduce the amount of data transmitted over the network in the literature [2]; Passive Clustering strategy is introduced into the Directed Diffusion Protocol in the Literatures [3, 4], plenty of reduction of the directional flooding in the interest and exploration of data, greatly improved the overall performance of the network.
When nodes propagate sensing data, the packet CH information field has five values: the unique marked node node_id, node_status, two CH nodes (ch1_id,ch2_id), energy_surplus(with a total of 17 bytes).
In other words, gateway only forward the packet of the CH which takes this node as CH; the ordinary nodes only receive packets from the CHs, and sent the data matching interest to the CH.
In a word, the proposed algorithm adopts CH declaration mechanism based on energy threshold and gateway selection mechanism based on residual energy, it has many merits: 1) it not only balances the energy consumption, but also reduces the probability of re-cluster's method in the network topology maintenance mechanism. 2) Comparing with PCDD ,EPCDD packet flooding overhead further decrease, and packet flooding overhead reduction is far greater than the overhead increase due to the increased length of the single information packet, so EPCDD has better energyefficiency; 3)the probability of collision greatly reduce while the number of network connections is further reduction, therefore, the packet's delay performance will be better, and packet reception rate is also higher;4) With the increasing of the nodes ,the greater is the network density, the more is the number of network packet, flooding overhead reduction and reduction in the number of the network connections become

The random variables of reliability analysis should be made statistical on the basis of test data.
Test data and statistical value are shown as table 1.
It shown that data can not be made statistical because there were only four of fault f212.
Failure probability is more than 99% if analyzed according to conventional data and index.
Slope engineering is usually designed under condition of without sufficient sample test data.