Search results

The abstraction of diagnostic feature from field condition monitoring data is a significant research challenge.
CCM Based Dimension Reduction 3.1 Dimension Reduction The degree of correlation between two parameters is represented by correlation coefficient, whereas CCM represents multi-parameter correlations.
(4)Set the correlation factor threshold T=0.80, and run the reduction steps in Matlab, the dimension is reduced to 9
The original feature vectors have 18 elements each before reduction, all of which belongs to three classes.
The test shows that after reduction the recognition rates are equal or even higher than that of before.

The selective reduction was promoted by selecting the appropriate amount of modifier.
Reduction order was elucidated in this paper, Fe was reduced from the slag followed by P, Mn and Si and the reduction rate of Si could reach about 51%.
The metal phase was rich in Fe, Si, Mn and P as a result of the selective reduction.
Effect of Temperature on Selective Reduction.
The variations in the recovery rate of Si are shown in Fig.4; here the SiO2 30 mass% in the slag was indicated by diamonds, circles represent the data points resulting from the reduction of 40 mass% SiO2 and the triangles denotes the 50 mass% SiO2 in the slag.

Principal Component Analysis (PCA) represents a powerful tool for analyzing data by reducing the number of dimensions, without important loss of information and has been applied on datasets in all scientific domains [4].
On the other hand, PCA is known as an unsupervised dimensionality reduction technique which transfers the data linearly and projects original data to a new set of parameters called the factors, while retaining as much as possible of the variation present in the data set.
So the first step in the synthesis of data comparison is dimensionless processing to the indicator's data.
The inverse indicator of the power plant data is shown in Table.2.
In our case study, we used five thermal power plant data; by means of PCA, we have got only ten factors that concentrate more than 60% of the information provided by the original five thermal power plant.

The synthesis processing of ZrB2 powders includes neutralizing precipitation and carborthermal reduction boronization.
In the present work, ZrB2 powders were synthesized based on the following carborthermal reduction boronization reaction (1) ZrO2 + 0.5B4C + 1.5C = ZrB2 + 2CO (g) (1) According to the thermodynamic data, we can calculate out the enthalpy of reaction (1) at standard condition to be ∆H298 = 589.568 kJ, indicating reaction (1) to be an endothermic reaction (thermodynamic data obtained from Ref.10).
This is due to incompletion of the carborthermal reduction boronization reaction at 1500o C.
Surface area (BET) was estimated using adsorption data in a relative pressure range from 0.05 to 0.1, and the result was shown in Table 1.
Hu: Handbook of Thermodynamic Data of Inorganics (Metallurgical Industry Publications, China, 2002,2nd edition) [11] H.

In order to test the effect of fast algorithm, the paper used several groups of data sets for comparison experiment.
The data sets used as shown in Table 1.
Fast Reduction Results.
The data in the decision table after making attribute reduction is used as three layer BP network training samples to train.
In Chinese [13] Jensen R,Shen Q.Semantics-preserving dimensionality reduction:rough and fuzzy- rough-based approaches.IEEE Transaction on Konowledge and Data Engineering,2004,16-(12):1457-1471

Remove the noise source data set and independent data, processing the missing data and clean dirty data, considering the time sequence and data changes.
In this paper, data cleaning work including missing values, noise data and inconsistent data processing.
The paper use FAP (Fill in with Average Poll result) [2] attribute structure on employment information, so as to evaluate analysis. 5) Data reduction.
Data reduction techniques can be used to obtain a reduced representation of data sets, although it is much smaller, but still close to maintain the integrity of the source data.
Data reduction method is adopted in this paper is: the data cube aggregation and dimensionality reduction. 6) Boolean transformation.

It is possible to pinpoint the actual location and shape of the cohesion zone in the blast furnace using experimental data on the temperatures of the beginning of softening and melting of iron ore materials.
Alekseev [15] first obtained experimental data on the temperatures of the beginning of softening and melting of iron ore materials, taking into account the degree of reduction.
Data were recorded in the range of angles from 5° to 90° in steps of 0.021° to 2θ and a point hold time of at least 2963 s.
The phase compositions and crystal structures of the samples were determined from X-ray diffraction data using the International Center for Diffraction Data (ICDD) PDF4 database.
Borisenko, Selection of the operating mode of BLT using the data on gas distribution on the grate and the shape of the melting zone in the lower part of the furnace, Metallurgist, 11 (2017) 40-46

It is recommended that the conductor should have a temperature-reduction value of 25℃ according to the conductor temperature-reduction value analyzed by amount of creep, which can be used for reference during line design and construction.
The software specially developed for a creep test is used to collect data in this test.
Data Analysis Method The relationship between creep elongation and time of stranded conductor changes exponential, and it is a linear relation on log-log coordinates.
Analyze test data to get the coefficient in creep equation; use the creep equation to calculate amount of conductor creep for 10 years or 30 years operation, and then offer a regression curve[7].
It can see temperature-reduction value of 5 types of conductors from Table 3.

The studied adsorption data fitted well to Langmuir adsorption model with the correlation coefficient 0.9947.
In the present investigation, the experimental data were tested with respect to the Langmuir isotherm.
The data obtained from the adsorption experiment conducted during the present investigation was fitted using different COD concentration into the isotherm equation.
The studied adsorption data fitted well to Langmuir adsorption model with the correlation coefficient 0.9947 in Fig. 6.
The studied adsorption data fitted well to Langmuir adsorption model with the correlation coefficient 0.9947.

First, the pre-system uses principal component model to convert properties of the source data of the basic window, and it plays a role of dimensionality reduction; Second, the post-system uses the density model to execute clustering operation; Finally, it uses the summary of data, generated in before two steps, to execute simply second clustering and update the clustering result, which fit the requirements of streaming data features.
The Description of a Dynamic Data Stream Model Data Stream.
,an), as the data come constantly , the old data are out from one end of the window and the new data enter from the other one end.
Dimension reduction has been one important research topic in the field of pattern recognition, machine learning, and multivariate data analysis.
Especially with the arrival of the information age, people get large amounts of data, high dimension, unstructured data has become easy increasingly, which makes data reduction has become more urgent.