Papers by Keyword: AIC

Abstract: Regularization item, subject to data misfit is used to stabilize the inversion problem. In selection of the trade-off parameter the Akaiche Information Criterion (AIC) has been applied to and compared to the chi-square misfit criterion. For the AIC we can achieve the trade-off parameter more conveniently by minimizing the AIC fuction. The result shows the AIC works well with an agreement to one accomplished from the Chi-square method. The result is very meaningful about applied technology of ocean science and geophysics.

Abstract: Basis on the AIC selection smallest regularization, flattest regularization and smoothest regularization resolution result are compared at different noise level. The result explore the insight into how to solve the inversion problem when noise increase gradually. Some analysis has been done to decide the ability of the three methods on handling the data gaps. Each method has their special way to invert the data and give inference for data gaps with zero padding, linearize interpolation or smooth way. The results of the study is very important ,especially in the applied technology problem such as geophysics, geodetics, and so on.

Abstract: Model selection procedures play important role in many researches especially quantitative research. . In several area of sciences, the analysis and model selection of experiments are often used and often contains two fundamental goals associated with the experimental response of interest which are to determine the best model. The way to address these goals is to implement a model selection procedure. Then, the objectives of this research are to determine whether or not the final models selected are in agreement or differ substantially across the three approaches to model selection: using Akaike’s Information Criterion, using a p-value criterion, and using a stepwise procedure.. Generally, results from these three models are usually compare to each other. All selected models are based on the heredity principle to design the possible model for each design. The actual data from literature, consisting of the 2x3 and 32 and 3x4 factorial designs are used to determine the final model. The results show that the P-Value WH and Stepwise methods give the highest percentage of matched model.

Abstract: To improve the multi-target detection performance of sensor arrays at low signal-to-noise ratios, a new method called DTAE（Detection Technique based on Approximate Eigenvectors）is proposed. In the proposed method, the approximate eigenvectors of the signal subspace corresponding to the estimate of the source cluster centroid are computed first, then the array output data are weighted by the approximate eigenvectors, finally the estimate of the number of signals is acquired by a certain criterion using the information of the peak-to-average power ratio of the weighted data in frequency domain. Simulations show DTAE demonstrates better detection performance than traditional methods such as EDT(Eigenvector Detection Technique) and AIC（Akaike Information Criterion）etc. at low SNR.

Abstract: In order to diagnose the health state of Aircraft effectively, a new method based on ARMA Model and probabilistic neural network(PNN) is proposed in this paper. First, an ARMA model is built using the original acoustic emission signal of aircraft crucial components, then use the autoregressive approximation theory to estimate model parameters, and order of the model is calculated according to Akaike Information Criterion(AIC). Use the autoregressive parameters to build feature vectors, then the probabilistic neural network is used to carry out the recognition of these feature vectors, and the health state of aircraft crucial components is effectively diagnosed. After the application on certain type of real aircraft, this method is proved to be capable of detecting the fatigue crack on crucial structural components. And we can conclude that the method is an effective way to carry out aircraft health diagnosis.

Abstract: Diagnosis of rolling elements bearing plays an important role on the running and maintenance of mechanical equipments. To enhance the feature of fault and to further diagnose the status of bearings with a small fault size so as to realize the early recognition, the method of inverse filter based on Autoregressive model is presented in this paper, and the corresponding criterion of order selection is also discussed. Analysis of simulation signals and real data show that this method could enhance feature of impulse signal. Meanwhile, it is also found that for small size fault, the root mean square feature is more effective than kurtosis value, which is considered very useful for early diagnosis of rolling elements bearings.

Abstract: X-ray analysis on iron ores and reduced iron powders revealed that around 60% acidinsoluble substances were hexagonal and tetragonal quartz, another 40% substances were sillimanite, alumina-silicate, an unnamed zeolite, all contained Si and Al. SEM images displayed that the particle size of them was in the range of 3~7 μm, which may be the initial source of the cracking in the sintered body. Statistics analysis showed that the Acid-Insoluble Content (AIC) for high-grade magnetite powder was (0.130±0.010) % during the latest five months. The predicting value for reduced iron powder from ore powders should be 0.179 %. However, the testing value for reduced iron powder was (0.192±0.014) %. The limited difference of 0.013% might imply rare pollution coming from the reduction and milling processes. The most important step for control AIC should be the separation process of iron ore powders.