Search results

Training data and test data of BP neural network had been reduced by rough set.
Training data and test data of BP neural network had been reduced by rough set.
Discretization data sample are shown in table 2, and Testing Data sample are shown in table 3 [9].
Training model was built with BP neural network, and data sample in table 1 was used as training data, and data sample in table 3 was used as testing data.
Rough set can simply data sample and it can save storage space, and it can avoid of data explosion.

Check the data.
Check data integrity, for example, the integral correlation data, the correct amount of data.
The amount of the source data is equal to the theoretical quantity of data and a washing out of the processing data.
Compare the data.
A large reduction is in the test cycle with the use of automated tool in data comparison, while it can improve the coverage of test data, and even up to 100% coverage, reduce the technical requirements for testers, and the staff freed from the single and complex work.

One of a challenging problem is how to deal with the data which shown the characteristics of high-dimensional, non-linear and multi-faceted nature.
How to effectively determine the appropriate neighborhood parameters are of great significance to obtain the correct low-dimensional structures for data dimensionality reduction and classification.
Application experiment analysis In order to verify the effectiveness of the proposed approach, we use standard data set of UCI Iris data [4] (sample points N = 150, dimension D = 4), making experimental analysis of mapping error.
The experiment shows that the category labels in the data handling to obtain the optimal adaptive neighborhood parameters, to benefit the purpose of data classification.
Laplacian eigenmaps for dimensionality reduction and data representation.

The main focus of this work was to fabricate Pt/porous n-Si gas sensors with a significant reduction in their breakdown voltages, Vbr.
An increase in electric field can cause a major reduction in the breakdown voltage Vbr.
Our experimental data shows that this current density is not adequate to increase the pores' depth significantly.
The comparison of breakdown data is provided in Tables 1.
As shown in this table the reductions in the breakdown voltages of our sample are significant.

Theoretical analysis based on the measured data of sandstone samples shows that the real COPL will be obviously underestimated if the COPL computed by the traditional point-count method is far below CEPL and dissolution processes generated the additional porosity.
The correct understanding of the contribution of diagenetic processes to porosity reduction and permeability retention is quite important for the prediction and evaluation of reservior quality [1-6].
Accordingly, they could quantificationally evaluate the relative importance of the cementation and compactional processes to the reduction of the original porosity of hydrocarbon reserviors [5-6].
Based on the measured data of sandstone samples from [5], the deviation in COPL resulting from a DIPL of -5.0% ranges from 0.75% to 2.10%, which is far below that for CEPL (i.e. 0.4~1.6%) (Table 1).
All sample data come from [5].

., Ltd Shenyang 110000 aliguodapple@sina.com bhesenliang@126.com cliguo_d@sina.com Key words: refrigeration room, limited space, vibration damping, noise reduction, absorber Abstract.
In the final noise test, the noise of refrigeration room has been reduced from 68.3 dB(A) to 63.2 dB(A) and the vibration floor has been significantly reduced which has got a good vibration and noise reduction effect.
Measurement and analysis of raw data noise characteristics A supermarket refrigeration room at supermarkets upper, when the device is running, the work unit's noise impact on the supermarket staff and consumers is great, in the supermarket shoppers can clearly feel the vibration due to the heat pump unit so that the vibration generated noise floor.
Conclusion By controlling the vibration and noise reduction, sound pressure level measurements around the heat pump 8 measuring points, an average of 63.2dB.
Absorber, and for low-IF noise, noise reduction measures using the perforated plate; against high-frequency noise, noise reduction measures adopted absorber.

Based on the experimental data and literature as well as using the CFD tools, a model of light fuel oil combustion has been developed with an emphasis on nitric oxide formation.
The quantitative results obtained are comparable to the experimental data.
Considering the nitrogen-bearing fuel, the precursor to NO is HCN which is converted to NO (oxidation) or N2 (reduction) [16-18].
On the quantitative basis, the obtained results are comparable to the experimental data.
Gil, Influence of pressure on the rate of nitric oxide reduction by char, Combustion and Flame 126 (2001) 1602 – 1606

As the data source of data mining, the data must be huge, contains noise, fuzzy and incomplete.
But in general, data mining includes what the next few process [4]: Data Preparation: the data which data mining will be deal with from different data sources, and it has large volumes of data, complex structure, which mixed with the vacancy data, noise data and redundant data.
Data selection: some of the data in the data source doesn't make any sense to build model and discover patterns.
Data mining: using a variety of data mining methods to analyze the related data.
And studies have shown that when a 5% reduction in the loss of customers, the average value of each customer can increase by more than 25% -100% [9].

Enchanced Expression of Gene Data Mrs.
A.Clustering high-dimensional data.
Feature subset selection is used for data reduction by removing redundant dimension.Subspace clustering is based on observation that different subspaces may contain different meaningful clusters.It searches group of clusters within different subspaces of same data set.Three techniques for effective clustering of high-dimensional data are:-dimension-growth subspace clustering (CLIQUE),dimension-reduction projected clustering(PROCLUS) & frequent pattern-based clustering(pCluster).
Authomatic subspace clustering of high dimensional data for data mining applications.
Raghavan, “Automatic subspace clustering of high dimensional data for data mining applications”, Proc.

In this paper, the energy loss indicators reduction analysis method is proposed based on rough sets theory using the SIS data of power plant.
Data qcquisition should be prior to reduction while the completeness of the data shoulb be ensured.
To discretization the data.
Reduction.
These data are used as the original data to operate the energy loss indicators rough set analysis.