Search results

There are a lot of continuous attributes in data streams, and classic rough set can only deal with discrete data; therefore, we must discrete the continuous attributes, which directly affects the subsequent attributes’ reduction and ensemble model construction.
As the data flow, we delete the old data synchronously and make sure that data appearing at the window are the need for processing, so as to realize the dynamic updating of the data set.
The data set of 12 attributes obtained by the reduction of rough set trains neural network classifiers, and integrating the results of each classifier to the ensemble.
On the last data chunk, concept drifting happened suddenly, and moreover, the accuracy after reduction is also higher than that before reduction.
Concluding Methods of traditional data mining face two major challenges of data stream: unlimited data and concept drifting.

Evaluating the Reliability of Satellite Gravity Data for Disaster Risk Mapping in West Coast Aceh's Coal Mines Adib Adib1,a*, Radhiyullah Armi2,b, Sabrian Tri Anda2,c, Azwanda Azwanda2,d 1University of Teuku Umar, West Aceh, Indonesia 2University of Samudra, Langsa, Indonesia aadib@utu.ac.id, bradhiyullah@unsam.ac.id, csabriantrianda@unsam.ac.id, dazwanda@utu.ac.id Keywords: Satellite Gravity Data, Disaster Risk Reduction, FHD, SVD.
This integration aligns with recent advancements in geophysical data analysis, such as those discussed by Pohan et al. [8], which emphasize the importance of combining multiple data sources for robust risk assessment.
The spatial resolution for latitude and longitude is 1 minute per grid, with an accuracy of roughly 0.1 mGal for gravity data and 1 meter for elevation data.
Future efforts should focus on creating integrated geophysical models that combine multiple data sources to improve predictive capabilities and address emerging challenges in disaster risk reduction.
This access has been essential in enabling the research and data analyzing done.

Fault Location Principle based on RS-IA Data Mining Model Overview.
It finds the laws from a large number of data by analyzing each data, and expresses them in the understandable form as far as possible [13, 15, 16].
Fault Location Based On RS-IA Data Mining Model Fault location based on the RS-IA data mining model mainly uses RS to extract domain knowledge, and obtains correlation rules of input and output vectors [14].
Fault Location of Distribution Network based on RS-IA Data Mining Model When there is a fault in some region of power distribution network, the fault location rules obtained by the RS-IA data mining model.
Sun, Data mining technology and its application on power system, Autom.

Introduction In our modern society, the volume and complexity of the collected data is growing rapidly.
In order to extract and describe the useful information in those data, the data mining emerges as a research area.
Data mining is a useful tool and method to solve the volume and complexity of the collected data.
Pawlak [1] in 1982, which is one of a new tools to dealing with vagueness and granularity in information systems, and it is able to find the useful information from volume and complexity of the collected data, moreover, it can guide people's decision by analyzed those useful information.
So is the N-reduction of fuzzy covering C.

Reduction of Nanometric Magnetite Powder Lj.
The experimental data suggest that the reoxidation process proceeds in two stages, at various activation energies.
The mass loss (∆mT∼31%) measured after the reduction process at T=833 K is somewhat larger than the theoretically predicted one (∆mT=27,64%) for the reduction of magnetite to iron.
Assuming that the reaction is diffusion controlled, Jander's model was applied, and the experimental data were correlated with the equation (full lines in Fig. 4) [8]: [1-(1-α) 1/3 ] 2 = kτ
According to the experimental and calculated data, (see Fig.5), it seems that the (re)oxidation process has two distinct stages, at activation energies of 0.17 and 0.59 eV.

The data driven refers to realizing the data predict that appraisal, dispatch, monitoring, diagnosis, decision-making and optimization of the system and so on each kind of expectation function by using the controlled system online and the off-line data [4].
Fig. 1 The multi-source information fusing system functional block diagram based on data driven Data Level Information Fusion.
Suppose there is n sensor to carry on measurement to some object X, each sensor measured data number is m, and then the sampled data matrix is:
Supposes the system input includes n sample, each sample Xi is composed of m data, produces the training sample data matrix (omitted).
References [1] Waltz E, Llinas J.Multisensor Data Fusion[M].

Traffic flow sampled data is noisy and chaotic time series.
Traditional ESN is only suitable for no noise or low-noise processing data, while the data with high noise prediction result is unsatisfactory[2].In interference of complex high-intensity noise component, simply optimization for ESN model does not completely solve the problem of prediction accuracy , while the use of data smoothing processing methods is often more direct and effective than optimal ESN model itself .The use of signal processing method filters out the data noise component, makes the data which is reduced noise have a higher signal noise ratio(SNA) , then provides the necessary conditions for high-precision prediction[3,4] .
The definition of goodness of fit, according to the References [6]: (2) Where is the original traffic flow data, is traffic flow data after noise reduction, is the sampling point .The Smaller is, the higher similarity to traffic flow before and after noise reduction is.
Multi-State Threshold Method Multi-state Threshold Method Which Is Based On Wavelet Transform In the Wavelet-based noise reduction algorithm, without any kind of soft-threshold method can be fit for all types of data.
Applicable threshold selection rules need us to fully consider characteristics of the denoising data.

Rough set theory has some advantages in the processing of data and eliminating redundant information and dealing with uncertain information,so widely used in data preprocessing, attribute reduction and so on.
Applications Based on Rough Sets and decision tree Experimental Data Acquisition The experimental data is from the Nong'an survey data, provided by Agricultural Technology Promotion Center.
Data Mining Based on Rough Set and Decision Tree combination.
Rough set attribute reduction algorithm requires that data is discrete data, according to the soil data characteristics, the Entropy / mdl discrete algorithm is carried to do the data processing.
Data Mining Results.In 1400 Records, correctly classified data is 1298, the others are not.

The researcher revises the questionnaire, practices questionnaire survey, analyzes data and selects 685 university students by stratified cluster sampling.
Data obtained are analyzed by descriptive statistics, independent sample t test, One-way ANOVA and structural equation model.
Data Processing (1) Cognition of energy saving and carbon reduction There are 17 items of “cognition of energy saving and carbon reduction”.
Path analysis of cognition and attitude toward energy saving and carbon reduction Regarding samples collected by formal test and by SEM program Amos 17.0, this study probes into relationship between students’ cognition and attitude toward energy saving and carbon reduction in order to validate the fit between theoretical model and actual data.
,Multivariate Data Analysis (5th ed.) , Englewood Cliffs, NJ : Prentice-Hal（1998）

Introduction Z.Pawlak, a Poland mathematician, has put forward Rough Set Theory a kind of new mathematics data analysis theory in 1982.
So, Rough Set is widely used in machine study, knowledge finding, knowledge acquisition, decision analysis, data mining, pattern-recognition, and so on, So far, Rough Set Theory already becomes the focus-studied in international artificial intelligence.
In this paper, on the basis of rough set theory about data reduction, using granular computing for knowledge reduction.
Export: A attribute reduction.
Thus the reduction is.