Search results

Analysis of the Sample Data Main influencing factors [2][3] of end sulfur content are as follows: weight of high carbon ferrochrome hot metal, the intensity of oxygen supply, the intensity of nitrogen supply, weight of chrome ore, weight of limestone, weight of ferrosilicon, quantity of slag, the initial sulfur content, the initial carbon content, initial phosphorus content, temperature of hot metal, melting time and melt temperature.
In order to eliminate noise sample data, we calculate the arithmetic mean and standard deviation of the sample set.
In order to obtain a better network performance, two most basic precondition must be met: first, the training set and testing set need to use a typical sample data; second, testing and training sets are different.
Using all sample data, regression equation was established, which included h components, as shown in function (1): (1) was the error sum of squares of y, was the prediction value of sample data i, was the actual value of the sample data.
Research on Prediction model stimulation Use 120 furnaces of field measured data from Sinosteel Jilin Ferroalloys co., Ltd smelting medium-low carbon ferrochrome as samples, PSO-PLSBP prediction model of end sulfur content for training and prediction, 60 furnaces of samples for training, 30 furnaces for verification, the remaining 30 furnaces for prediction.

The discussed problems result the lack of scores or access records ,which means the collected data sets are often high dimensional and sparse, leading to the Dimension Reduction(DR) before using the traditional clustering techniques[7].
However, DR makes the noise data and valid data closer to each other in lower dimensional space and the loss of some important correlation, which decreases the clustering quality.
As for CS based on context, because of introduction of more context factors, the high-dimensionality and sparsity turn to be more serious, dataset capacity also becomes greater, so how to keep or improve the quality of CS based on preservation of original correlations between data items is a significant problem needed further research.
Finally 85 customers’ respective scores for random 20 comedy films and associated context data are collected, after the calculation of, awith 85 vertices and 5 hyper edges is constructed.
[2] M.Gorgoglione, C.Palmisano and A.Tuzhilin,in:Personalization in context: Does context matter when building personalized customer models,Proceedings of the Sixth International Conference on Data Mining, Boston(2006), in press

In order to obtain more accurate parameters estimate and compensate for defects in the least-squares method, the recursive algorithm is applied to the input data to make parameter estimates constantly revised.
According to the identification comparative analysis, the data acquired from the sensors that are close to magnetic vibration isolator have high model fitting degree, good correlation and truly reflect dynamic state of the system.
So we choose B12( the data acquired form the sensor 2 when vibration isolator 1 operates)and B25( the data acquired from the sensor5 when vibration isolator 2 operates) as the research objects.
Fig. 3 B12 frequency response fitting curve Fig. 4 B12 frequency response fitting curve of three sets of experimental data of three sets of experimental data Matlab are used to do the simulation analysis, set the system sampling frequency for 2048HZ, the bandpass filter rang from 10 to 200HZ, the simulation step for 12000,the bandpass filter order for 6, the B12 model order for 15, the B25 model order for 13 .
Magnetic active vibration isolation system as object, based on field data collected, we use the recursive least squares method for system identification to get the system model and transfer function matrix.

In the initial phase of works, the literature and Internet databases containing structural data of crystalline compounds were reviewed in search of a model of tridymite structure which best fits the one forming in silica materials.
The presence of quartz and cristobalite in silica materials is an unfavourable phenomenon due to low-temperature polymorphous changes of quartz  at 573C and cristobalite  at 230C to 180C, which are accompanied by a considerable change in volume, resulting in the formation of microcracks in the material and consequent reduction in the strength.
Crystallographic data for all the tested models may be found in Internet databases [4 and 5].
Summary  On the basis of the collected literature data, the structure of tridymite in refractory silica materials has been refined.

In a second step, the thermal maps extracted from thermographic monitoring are used as input data for thermal-mechanical calculations.
It was observed, in all cases, significant changes of the disc bainitic microstructure, i.e: martensitic transformation and a drastic grain-size reduction, on a depth of 4-6 mm.
In a second step, the thermalmechanical calculations, carried out using thermographic data as loading input, are presented.
A thermal calculation is first performed, using experimental infrared-video data as surface temperature field at each time, in order to determine at each time and in each point the bulk temperature.

According to the obtained data of the wetting angle, the regression equation was derived.
The significant limitations have low rates of adhesion changes, bulk polymer destruction - reduction of mechanical resistance to loads Chemical Treatment Chemical method of surface modification is the main method of modification of PCM in industry.
Based on the data obtained, a mathematical model of the dependence of the change in the contact properties of the resultant feature (wetting angle) of the PTFE film on the factor values of which are: voltage (x1) and discharge current (x2) in the reaction chamber.
According to the data obtained, a matrix of experimental planning of the wetting angle depends on 2 factors: amperage (I) and voltage (U) (table. 2).

Additionally, the isotherm experimental data fits the Langmuir adsorption isotherm more closely, and the maximum adsorption capacity of Cr(VI) on β-FeOOH and β-FeOOH/SMZ can attain 20.79 mg/g and 22.08 mg/g, respectively.
Among various methods in Cr (VI) removal such as reduction, precipitation, ion exchange and adsorption [2], adsorption is widely used for its low cost and high efficiency, and it is generally considered to be a promising method.
The isotherm experimental data shows good fit to both Freundlich and Langmuir isotherms but the latter is better in liner correlation which indicates the mechanism is monolayer adsorption.
Equilibrium adsorption data of adsorbents indicates well fit to Langmuir sorption isotherm suggesting the mechanism of monolayer adsorption, and the adsorption capacity for Cr (VI) on β-FeOOH and β-FeOOH/SMZ can achieve 20.79 mg/g and 22.08 mg/g, respectively.

The data points creating these diagrams were filtered first, subsequently processed and then evaluated using the double-K fracture model.
GTDiPS software [3] was used to edit the diagrams (elimination of data point duplication and reduction of the number of such points).
Application of the Double-K Fracture Model Input data for the double-K fracture model were selected from the above-mentioned edited Pv–CMOD diagrams.

The noise reduction coefficient (NRC) of sample D is 38.26% while for sample C is 37.42%.
But there are opposite result as shown in Fig. 4 which is foam thickness from D, E and F sound absorption increases but sometime fluctuating data found at the lower frequency.

The use of physical data mining data mining, it is easy to obtain meaningful results of the statistical methods are difficult to get.
First, make data preprocessing, data need to be analyzed to extract, remove no student achievement, this step data reduction from 6097 to 5513. 2. analyzed using Pearson correlation coefficient height and lung capacity, endurance, grip strength, jumping separate correlation results are as table 2.
Towards imp roving data quality [C ].
IEEE Data EngineeringBulletin, 2000, 23 (4) : 3- 13
Cleaning the spurious links in data [ J ].