AIC Methods Research Articles

Research on Seismic Phase Recognition Method Based on Bi-LSTM Network

In order to improve the precision of phase recognition and reduce the rate of misdetection, this paper applies the deep learning method to automatic phase recognition. In this paper, an automatic seismic phase recognition model based on the Bi-LSTM network is designed. To test the performance of this model, the STEAD dataset is used for training and testing, and this model is compared with the traditional STA/LTA and AIC methods. The experimental results show that, compared to STA/LTA and AIC methods, the Bi-LSTM network can reduce the misdetection rate by about 8–15%, and improve the RSEM; especially, the prediction error of S-wave is greatly reduced.

Determine the factors affecting the time to recovery of children with bacterial meningitis at Jigjiga university referral hospital in the Somali Regional State of Ethiopia: using the parametric shared frailty and AFT models

BackgroundNeisseria meningitides, Streptococcus pneumonia, and hemophilic influenza type B are frequently linked to bacterial meningitis (BM) in children. It’s an infectious sickness that kills and severely mobilizes children. For a variety of reasons, bacterial meningitis remains a global public health concern; most cases and deaths are found in Sub-Saharan Africa, particularly in Ethiopia. Even though vaccination has made BM more preventable, children worldwide are still severely harmed by this serious illness. Age, sex, and co-morbidity are among the risk variables for BM that have been found. Therefore, the main objective of this study was to identify the variables influencing the time to recovery for children with bacterial meningitis at Jigjiga University referral hospital in the Somali regional state of Ethiopia.MethodA retrospective cohort of 535 children with bacterial meningitis who received antibiotic treatment was the subject of this study. Parametric Shared Frailty ty and the AFT model were employed with log likelihood, BIC, and AIC methods of model selection. The frailty models all employed the patients' kebele as a clustering factor.ResultsThe number of cases of BM declined in young children during the duration of the 2 year, 11 month study period, but not in the elderly. Streptococcus pneumonia (50%), hemophilic influenza (30.5%), and Neisseria meningitides (15%) were the most frequent causes of BM. The time to recovery of patients from bacteria was significantly influenced by the covariates male patients (ϕ = 0.927; 95% CI (0.866, 0.984); p-value = 0.014), patients without a vaccination history (ϕ = 0.898; 95% CI (0.834, 0.965); P value = 0.0037), and patients who were not breastfeeding (ϕ = 0.616; 95% CI (0.404, 0.039); P-value = 0.024). The recovery times for male, non-breastfed children with bacterial patients are 7.9 and 48.4% shorter, respectively. In contrast to children with comorbidity, the recovery time for children without comorbidity increased by 8.7%.ConclusionAge group, sex, vaccination status, co-morbidity, breastfeeding, and medication regimen were the main determinant factors for the time to recovery of patients with bacterial meningitis. Patients with co-morbidities require the doctor at Jigjiga University Referral Hospital to pay close attention to them.

Bayesian model averaging for predicting factors associated with length of COVID-19 hospitalization

IntroductionThe length of hospital stay (LOHS) caused by COVID-19 has imposed a financial burden, and cost on the healthcare service system and a high psychological burden on patients and health workers. The purpose of this study is to adopt the Bayesian model averaging (BMA) based on linear regression models and to determine the predictors of the LOHS of COVID-19.MethodsIn this historical cohort study, from 5100 COVID-19 patients who had registered in the hospital database, 4996 patients were eligible to enter the study. The data included demographic, clinical, biomarkers, and LOHS. Factors affecting the LOHS were fitted in six models, including the stepwise method, AIC, BIC in classical linear regression models, two BMA using Occam's Window and Markov Chain Monte Carlo (MCMC) methods, and GBDT algorithm, a new method of machine learning.ResultsThe average length of hospitalization was 6.7 ± 5.7 days. In fitting classical linear models, both stepwise and AIC methods (R2 = 0.168 and adjusted R2 = 0.165) performed better than BIC (R2 = 0.160 and adjusted = 0.158). In fitting the BMA, Occam's Window model has performed better than MCMC with R2 = 0.174. The GBDT method with the value of R2 = 0.64, has performed worse than the BMA in the testing dataset but not in the training dataset. Based on the six fitted models, hospitalized in ICU, respiratory distress, age, diabetes, CRP, PO2, WBC, AST, BUN, and NLR were associated significantly with predicting LOHS of COVID-19.ConclusionThe BMA with Occam's Window method has a better fit and better performance in predicting affecting factors on the LOHS in the testing dataset than other models.

A robust method for coherent and non-coherent source number detection using a special Hankel-based covariance matrix

A robust algorithm for source number estimation based on the formation of the Hankel covariance matrix is presented. First, multiple data snapshots are taken successively from overlapped subarrays in a way similar to the forward spatial smoothing method to construct the special Hankel covariance matrix and for the total number of subarrays, these special covariance matrices are generated. Then, the average of these matrices is employed in singular value decomposition to generate the corresponding eigenvalues. Finally, the resulting eigenvalues are evaluated via the rule presented in this paper as the Moving Gradient Criterion (MGC) to estimate the number of sources by detection of the largest singular values. The greatest difference between the proposed algorithm and the other conventional methods is the form of the covariance matrix with the observed signal that can handle both non-coherent as well as fully coherent sources. Also, the proposed MGC rule adopted with this form of the covariance matrix is the strength of this work. Numerical simulations demonstrate the high superiority of the proposed approach over the competing methods such as MDL, AIC, SORTE, RAE and MSEE methods, especially in the cases of very closely spaced sources, low SNR values, low sensors number and low snapshots number.

Quantitative CT imaging features for COVID-19 evaluation: The ability to differentiate COVID-19 from non- COVID-19 (highly suspected) pneumonia patients during the epidemic period.

COVID-19 and Non-Covid-19 (NC) Pneumonia encountered high CT imaging overlaps during pandemic. The study aims to evaluate the effectiveness of image-based quantitative CT features in discriminating COVID-19 from NC Pneumonia. 145 patients with highly suspected COVID-19 were retrospectively enrolled from four centers in Sichuan Province during January 23 to March 23, 2020. 88 cases were confirmed as COVID-19, and 57 patients were NC. The dataset was randomly divided by 3:2 into training and testing sets. The quantitative CT radiomics features were extracted and screened sequentially by correlation analysis, Mann-Whitney U test, the least absolute shrinkage and selection operator (LASSO) logistic regression (LR) and backward stepwise LR with minimum AIC methods. The selected features were used to construct the LR model for differentiating COVID-19 from NC. Meanwhile, the differentiation performance of traditional quantitative CT features such as lesion volume ratio, ground glass opacity (GGO) or consolidation volume ratio were also considered and compared with Radiomics-based method. The receiver operating characteristic curve (ROC) analysis were conducted to evaluate the predicting performance. Compared with traditional CT quantitative features, radiomics features performed best with the highest Area Under Curve (AUC), sensitivity, specificity and accuracy in the training (0.994, 0.942, 1.0 and 0.965) and testing sets (0.977, 0.944, 0.870, 0.915) (Delong test, P < 0.001). Among CT volume-ratio based models using lesion or GGO component ratio, the model combining CT lesion score and component ratio performed better than others, with the AUC, sensitivity, specificity and accuracy of 0.84, 0.692, 0.853, 0.756 in the training set and 0.779, 0.667, 0.826, 0.729 in the testing set. The significant difference of the most selected wavelet transformed radiomics features between COVID-19 and NC might well reflect the CT signs. The differentiation between COVID-19 and NC could be well improved by using radiomics features, compared with traditional CT quantitative values.

Spatial Distributions and Model Selections of Commercial Estuarine Fish (Sciaenidae) Populations Related to Water Quality, Chl-a, and AML in Musi River mouth, South Sumatra

Estuary and river mouth are essential habitats for many commercial estuarine fishes, including the Sciaenidae family. While recently, estuaries have been threatened by anthropogenic marine litter (AML) transported from nearby land and river. An important type of AML is plastic litter since it takes a long degradation time. In the South Sumatra Province, Indonesia, one of the vital estuaries is the Musi estuary. This paper aims to map the spatial distributions of two Sciaenids, including Panna microdon and Otolithoides pama, and Sciaenid’s environmental covariates, including water quality, chlorophyll a, and plastic litters in Musi estuary and model the correlations of Sciaenids with their covariates. The maps were developed using GIS, and the model was validated using AIC methods. The data were collected from 3 river mouths in the west, central, and east of the Musi estuary. The data showed that the populations of both Sciaenids were higher in the east river mouth rather than in the west. Sciaenid populations were positively correlated with high salinity, DO, chlorophyll a, moderate transparency, and low temperature. A high load of AML’s frequency (7.54 items/m2) and weights (36.8 gram/m2) has reduced both Sciaenid populations in the central river mouth of the estuary. In contrast, low AML loads in the east have correlated with high Sciaenid populations. Model selection based on AIC values shows the best model for P.microdon retained an effect of AML weight with AIC values of 22.591 and 28.321 for O. pama. This concludes that the weight of plastic litter in estuary water was the main limiting factor for Sciaenid populations in Musi.

A neural network for determination of latent dimensionality in non-negative matrix factorization

Non-negative matrix factorization (NMF) has proven to be a powerful unsupervised learning method for uncovering hidden features in complex and noisy data sets with applications in data mining, text recognition, dimension reduction, face recognition, anomaly detection, blind source separation, and many other fields. An important input for NMF is the latent dimensionality of the data, that is, the number of hidden features, K, present in the explored data set. Unfortunately, this quantity is rarely known a priori. The existing methods for determining latent dimensionality, such as automatic relevance determination (ARD), are mostly heuristic and utilize different characteristics to estimate the number of hidden features. However, all of them require human presence to make a final determination of K. Here we utilize a supervised machine learning approach in combination with a recent method for model determination, called NMFk, to determine the number of hidden features automatically. NMFk performs a set of NMF simulations on an ensemble of matrices, obtained by bootstrapping the initial data set, and determines which K produces stable groups of latent features that reconstruct the initial data set well. We then train a multi-layer perceptron (MLP) classifier network to determine the correct number of latent features utilizing the statistics and characteristics of the NMF solutions, obtained from NMFk. In order to train the MLP classifier, a training set of 58 660 matrices with predetermined latent features were factorized with NMFk. The MLP classifier in conjunction with NMFk maintains a greater than 95% success rate when applied to a held out test set. Additionally, when applied to two well-known benchmark data sets, the swimmer and MIT face data, NMFk/MLP correctly recovered the established number of hidden features. Finally, we compared the accuracy of our method to ARD, AIC and stability-based methods.

Comparison of variable selection methods for clinical predictive modeling

ObjectiveModern machine learning-based modeling methods are increasingly applied to clinical problems. One such application is in variable selection methods for predictive modeling. However, there is limited research comparing the performance of classic and modern for variable selection in clinical datasets. Materials and MethodsWe analyzed the performance of eight different variable selection methods: four regression-based methods (stepwise backward selection using p-value and AIC, Least Absolute Shrinkage and Selection Operator, and Elastic Net) and four tree-based methods (Variable Selection Using Random Forest, Regularized Random Forests, Boruta, and Gradient Boosted Feature Selection). We used two clinical datasets of different sizes, a multicenter adult clinical deterioration cohort and a single center pediatric acute kidney injury cohort. Method evaluation included measures of parsimony, variable importance, and discrimination. ResultsIn the large, multicenter dataset, the modern tree-based Variable Selection Using Random Forest and the Gradient Boosted Feature Selection methods achieved the best parsimony. In the smaller, single-center dataset, the classic regression-based stepwise backward selection using p-value and AIC methods achieved the best parsimony. In both datasets, variable selection tended to decrease the accuracy of the random forest models and increase the accuracy of logistic regression models. ConclusionsThe performance of classic regression-based and modern tree-based variable selection methods is associated with the size of the clinical dataset used. Classic regression-based variable selection methods seem to achieve better parsimony in clinical prediction problems in smaller datasets while modern tree-based methods perform better in larger datasets.

Modelling the non-isothermal crystallization of polymers: Application to poly(ethylene 2,5-furandicarboxylate)

Several equations have been compared for the modelling of semi-cristalline polymer crystallization on cooling from the melt and on heating from the glass. The temperature dependence was described with the Hoffman–Lauritzen (HL) equation, where U* and Kg were first determined by fitting the activation energy (Eα) dependency with temperature computed with an advanced isoconversional method (AIC) from DSC data. Several models were used to describe the extent of crystallization dependence, such as Sestak-Berggren and Avrami models. Simulations using Ozawa and AIC methods were performed for comparisons. It was shown that the model-free approach is able to take into account additional crystallization phenomena that are not considered in the classical HL or Avrami equations. Then, a new equation has been proposed to simulate additional phenomena occurring at the end of crystallization on heating from the glass, which are not described by the Hoffman–Lauritzen theory. The experimental finding that application of an isoconversional method to crystallization data following a HL mechanism, should lead to positive decreasing values for crystallization from the glass and to negative increasing values for crystallization from the melt was confirmed by simulations.

더러브렛 경주마의 주파기록에 대한 유전모수 추정과 연도별 개량량 분석

본 연구는 1990년부터 2008년까지 서울경마공원과 부산 경마공원에서 활동한 경주마 13,458두의 총 245,979개의 경주관측기록을 이용하여 국내 더러브렛 경주마의 유전능력과 연도별 개량량을 분석하였다. 모든 관측치들의 주파기록을 1400m 경주거리에 맞추어 보정하고, 기수와 조교사에 대해 평균수득상금의 순위에 따라 번호를 부여하였으며, 후보 설명변수들에 대해 AIC, BIC를 사용한 단계적 변수 선발법을 적용하였다. 19년간 국내에서 활동한 경주마들의 주파기록에 대한 유전력과 반복력은 각각 0.322, 0.332로 추정되었고, 연평균 표현형 개량량과 유전적 개량량은 -0.166초, -0.161초로 산출되었다. 경주마의 유전모수와 개량량의 추정 연구에 있어 무엇보다 설명력 높은 개체모형의 수립이 중요하다고 생각되며, 이는 실제 경주마의 능력을 효과적으로 반영하는 반응변수의 설정과 통계학적 기법을 활용한 체계적인 설명변수의 선발을 통해 가능하다 The aim of this study was to estimate genetic parameters and annual trends on the racing performance of Thoroughbred horses by a statistical analysis of the resulting records. We used the racing results of 245,979 observations for 13,458 horses recorded in 19 years of race held at Seoul and Busan racing tracks, provided by Korea Racing Authority. After a careful adjustment of some variables such as racing times, jockey and trainer numbers and the average prize a horse won, we selected significant factors that explain the result of racing records of a horse by stepwise AIC and BIC methods. The estimated heritability and repeatability were 0.322 and 0.332, respectively. The average of annual phenotypic and genetic improvement was -0.166 seconds and -0.161 seconds, respectively. Based on the statistical approach, we established reasonable animal model of well-set variables, which is important in the study on estimating performance of racing horses.

Accurate Source Number Detection Using Pre-Estimated Signal Subspace

This paper presents a scheme for accurately detecting the number of incident waves arriving at array antennas. The array antenna and MIMO techniques are developing as 4th generation mobile communication systems and wireless LAN technologies, and the accurate estimation of the propagation environment is becoming more important. This paper emphasizes the accurate detection of the number of incident waves; one of the important characteristics in multidirectional communication. There are some recent papers on accurate detection but they have problems of estimation accuracy or computational cost in severe environment like low SNR, small number of snapshots or waves with close angles. Hence, AIC and MDL methods based on statistics and information theory are still often used. In this paper, we propose an accurate estimation method of the number of arrival signals using the orthogonality of subspaces derived from preliminary estimation of signal subspace. The proposed method accurately estimates the number of signals also in severe environments where AIC and MDL methods can hardly estimate. We evaluate the performance of these methods through some computer simulation and experiments in anechoic chamber.

Estimating the order of a hidden markov model

AbstractWhile the estimation of the parameters of a hidden Markov model has been studied extensively, the consistent estimation of the number of hidden states is still an unsolved problem. The AIC and BIC methods are used most commonly, but their use in this context has not been justified theoretically. The author shows that for many common models, the penalized minimum‐distance method yields a consistent estimate of the number of hidden states in a stationary hidden Markov model. In addition to addressing the identifiability issues, she applies her method to a multiple sclerosis data set and assesses its performance via simulation.

Estimation of the order of an auto-regressive model

There are no definitive solutions available for the order estimation of the auto-regressive process. First, the performance of the three criteria, namely FPE, AIC and MDL is illustrated. Next, it is indicated that there are possibilities of their performance being improved. The algorithmtri proposed here utilizes three minimum values of any of the conventional loss functions in the FPE, AIC and MDL methods. It also uses three statistics derived from these three minimum values. The estimated order is a rounded weighted average of these six statistics. The algorithm is found to do better in a qualified sense of yielding peakier distribution of the estimated orders when tested for 1000 synthetic models of orders 3, 5 and 7 each. The conclusion drawn is that there are open possibilities of improving upon the conventional order estimators for auto-regressive processes. This means that till axiologically sounder estimators are available one should be ready to use heuristic estimators proposed here.

A distribution-free technique to estimate the order of markovian dependence using entropy

A model-order determination procedure for time series is described which is based on computing the differential entropy of the time series' multivariate amplitude probability density. The procedure, termed the differential entropy method, applies to linear and nonlinear models for the time series. Simulations were performed to determine the performance characteristics of the differential entropy method on first-order autoregressive models with Gaussian and non-Gaussian inputs and on a first-order nonlinear model. It produces accurate model-order estimates for the linear model once the correlation coefficient becomes large enough. The threshold for accurate estimates is dependent on the amplitude distribution of the model's input. This technique estimates the model order of data having nonlinear dependence structure more accurately than thead hoc application of the AIC and MDL methods. Extension of the method to higher-order models is described, but, because of computation complexity issues, the differential entropy method is best used for estimating small model orders.

An Optimal Selection of Regression Variables

An asymptotically optimal selection of regression variables is proposed. The key assumption is that the number of control variables is infinite or increases with the sample size. It is also shown that Mallows's Cp', Akaike's FPE and aic methods are all asymptotically equivalent to this method.

An optimal selection of regression variables

SUMMARY An asymptotically optimal selection of regression variables is proposed. The key assumption is that the number of control variables is infinite or increases with the sample size. It is also shown that Mallows's Qp, Akaike's FPE -and AIC methods are all asymptotically equivalent to this method.