• Corrosion Science and Technology
• /
• v.19 no.6
• /
• pp.302-309
• /
• 2020

This study aimed to predict the pitting corrosion characteristics of AL-6XN super-austenitic steel using multiple linear regression. The variables used in the model are degree of sensitization, temperature, and pH. Experiments were designed and cyclic polarization curve tests were conducted accordingly. The data obtained from the cyclic polarization curve tests were used as training data for the multiple linear regression model. The significance of each factor in the response (critical pitting potential, repassivation potential) was analyzed. The multiple linear regression model was validated using experimental conditions that were not included in the training data. As a result, the degree of sensitization showed a greater effect than the other variables. Multiple linear regression showed poor performance for prediction of repassivation potential. On the other hand, the model showed a considerable degree of predictive performance for critical pitting potential. The coefficient of determination (R2) was 0.7745. The possibility for pitting potential prediction was confirmed using multiple linear regression.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2011.05a
• /
• pp.509-511
• /
• 2011

본 논문에서는 다층 퍼셉트론 신경회로망과 선형예측계수를 사용한 화자인식 알고리즘을 제안한다. 제안하는 화자인식 알고리즘은 입력받은 음성신호에 대해서 유성음 구간을 추출한다. 추출된 유성음구간에 대하여 선형예측 분석에 의하여 화자의 특성을 가지고 있는 선형예측계수를 구한다. 구해진 선형예측계수를 분류하기 위하여 선형예측계수를 퍼셉트론 신경회로망의 입력으로 사용하여 네트워크의 학습을 수행한다. 본 실험에서는 선형예측계수와 신경회로망을 사용하여 본 화자인식 알고리즘이 유효하다는 것을 인식률을 통하여 확인한다.

• Proceedings of the IEEK Conference
• /
• 2000.07a
• /
• pp.260-263
• /
• 2000

A technique that uses linear prediction to achieve noise reduction in a voice signal which has been mixed with an ambient noise (Signal to Noise (S-N) ratio = about 0dB) is proposed. This noise reduction method which is based on the linear prediction estimates the voice spectrum while ignoring the spectrum of the noise. The performance of the noise reduction method is first examined using the transversal linear predictor filter. However, with this method there is deterioration in the tone quality of the predicted voice due to the low level of the S-N ratio. An additional processing circuit is then proposed so as to adjust the noise reduction circuit with an aim of improving the problem of tone deterioration. Next, we consider a practical application where the effects of round on errors arising from fixed-point computation has to be minimized. This minimization is achieved by using the lattice predictor filter which in comparison to the transversal type, is Down to be less sensitive to the round-off error associated with finite word length operations. Finally, we consider a practical application where noise reduction is necessary. In this noise reduction method, both the voice spectrum and the actual noise spectrum are estimated. Noise reduction is achieved by using the linear predictor filter which includes the control of the predictor filter coefficient’s update.

• International Journal of Advanced Culture Technology
• /
• v.11 no.3
• /
• pp.310-314
• /
• 2023

The purpose of this study is to predict the remaining capacity of lithium-ion batteries and evaluate their performance using five artificial intelligence models, including linear regression analysis, decision tree, random forest, neural network, and ensemble model. We is in the study, measured Excel data from the CS2 lithium-ion battery was used, and the prediction accuracy of the model was measured using evaluation indicators such as mean square error, mean absolute error, coefficient of determination, and root mean square error. As a result of this study, the Root Mean Square Error(RMSE) of the linear regression model was 0.045, the decision tree model was 0.038, the random forest model was 0.034, the neural network model was 0.032, and the ensemble model was 0.030. The ensemble model had the best prediction performance, with the neural network model taking second place. The decision tree model and random forest model also performed quite well, and the linear regression model showed poor prediction performance compared to other models. Therefore, through this study, ensemble models and neural network models are most suitable for predicting the remaining capacity of lithium-ion batteries, and decision tree and random forest models also showed good performance. Linear regression models showed relatively poor predictive performance. Therefore, it was concluded that it is appropriate to prioritize ensemble models and neural network models in order to improve the efficiency of battery management and energy systems.

• Journal of Korean Association for Spatial Structures
• /
• v.23 no.3
• /
• pp.87-94
• /
• 2023

Chloride is one of the most common threats to reinforced concrete (RC) durability. Alkaline environment of concrete makes a passive layer on the surface of reinforcement bars that prevents the bar from corrosion. However, when the chloride concentration amount at the reinforcement bar reaches a certain level, deterioration of the passive protection layer occurs, causing corrosion and ultimately reducing the structure's safety and durability. Therefore, understanding the chloride diffusion and its prediction are important to evaluate the safety and durability of RC structure. In this study, the chloride diffusion coefficient is predicted by machine learning techniques. Various machine learning techniques such as multiple linear regression, decision tree, random forest, support vector machine, artificial neural networks, extreme gradient boosting annd k-nearest neighbor were used and accuracy of there models were compared. In order to evaluate the accuracy, root mean square error (RMSE), mean square error (MSE), mean absolute error (MAE) and coefficient of determination (R2) were used as prediction performance indices. The k-fold cross-validation procedure was used to estimate the performance of machine learning models when making predictions on data not used during training. Grid search was applied to hyperparameter optimization. It has been shown from numerical simulation that ensemble learning methods such as random forest and extreme gradient boosting successfully predicted the chloride diffusion coefficient and artificial neural networks also provided accurate result.

• Asian-Australasian Journal of Animal Sciences
• /
• v.29 no.1
• /
• pp.36-42
• /
• 2016

Milk-related traits (milk yield, fat and protein) have been crucial to selection of Holstein. It is essential to find the current selection trends of Holstein. Despite this, uncovering the current trends of selection have been ignored in previous studies. We suggest a new formula to detect the current selection trends based on single nucleotide polymorphisms (SNP). This suggestion is based on the best linear unbiased prediction (BLUP) and the Fisher's fundamental theorem of natural selection both of which are trait-dependent. Fisher's theorem links the additive genetic variance to the selection coefficient. For Holstein milk production traits, we estimated the additive genetic variance using SNP effect from BLUP and selection coefficients based on genetic variance to search highly selective SNPs. Through these processes, we identified significantly selective SNPs. The number of genes containing highly selective SNPs with p-value <0.01 (nearly top 1% SNPs) in all traits and p-value <0.001 (nearly top 0.1%) in any traits was 14. They are phosphodiesterase 4B (PDE4B), serine/threonine kinase 40 (STK40), collagen, type XI, alpha 1 (COL11A1), ephrin-A1 (EFNA1), netrin 4 (NTN4), neuron specific gene family member 1 (NSG1), estrogen receptor 1 (ESR1), neurexin 3 (NRXN3), spectrin, beta, non-erythrocytic 1 (SPTBN1), ADP-ribosylation factor interacting protein 1 (ARFIP1), mutL homolog 1 (MLH1), transmembrane channel-like 7 (TMC7), carboxypeptidase X, member 2 (CPXM2) and ADAM metallopeptidase domain 12 (ADAM12). These genes may be important for future artificial selection trends. Also, we found that the SNP effect predicted from BLUP was the key factor to determine the expected current selection coefficient of SNP. Under Hardy-Weinberg equilibrium of SNP markers in current generation, the selection coefficient is equivalent to $2^*SNP$ effect.

• Journal of the Korean earth science society
• /
• v.32 no.6
• /
• pp.629-639
• /
• 2011

Based on Maxwell-Eucken(ME) model, which is one of structural models, a new model for predicting the effective thermal conductivity of variably saturated porous materials is proposed. The new model is a linear combination of three ME models having matrix, water, and air as a continuous phase. The coefficient of the corresponding linear equation is defined by a parameter referred to as 'the continuity coefficient', which provides a relative degree of continuity of each phase. The continuity coefficient of matrix is assumed to be linearly proportional to porosity. The model can be linear or nonlinear depending on how the continuity coefficients of water and air vary with water saturation. The feasibility of the proposed model was examined by both numerical and experimental results. Both linear and nonlinear models showed a high accuracy of prediction with $R^2$ values of 0.86-0.98 and 0.88-0.99, respectively. The numerical and experimental results also showed that the continuity coefficient of matrix was linearly proportional to porosity. Therefore, the proposed prediction model can be effectively used to estimate effective thermal conductivity of unsaturated porous materials by measuring porosity, water content and mineralogical compositions of matrix.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.19 no.4
• /
• pp.1-5
• /
• 1982

A new algorithm, based upon a lattice formulation, is presented for linear prediction. The output of the algorithm is the reflection coefficients that guarantee the stability of the all-pole model. The equations are derived that compute the covariance of the residuals recursively at each prediction stage, and in processing of computing that eqations, the reflection coefficients are estimated without computing the predictor coefficients. Comparing with covariance-lattice method, it can be said that the new algorithm reduce the number of computations to about half and is more efficient for fitting of the high-order model.

• The Journal of the Acoustical Society of Korea
• /
• v.24 no.1E
• /
• pp.16-20
• /
• 2005

A novel scheme to measure the speaker information in speech signal is proposed. We develope the theory of quantitative measurement of the speaker characteristics in the information theoretic point of view, and connect it to the classification error rate. Homomorphic analysis based features, such as mel frequency cepstral coefficient (MFCC), linear prediction cepstral coefficient (LPCC), and linear frequency cepstral coefficient (LFCC) are studied to measure speaker specific information contained in those feature sets by computing mutual information. Theories and experimental results provide us quantitative measure of speaker information in speech signal.

• Phonetics and Speech Sciences
• /
• v.7 no.4
• /
• pp.9-16
• /
• 2015

In this paper, several experiments are performed according to deep neural network (DNN) based features for the performance comparison of speaker verification (SV) systems. To this end, input features for a DNN, such as mel-frequency cepstral coefficient (MFCC), linear-frequency cepstral coefficient (LFCC), and perceptual linear prediction (PLP), are first compared in a view of the SV performance. After that, the effect of a DNN training method and a structure of hidden layers of DNNs on the SV performance is investigated depending on the type of features. The performance of an SV system is then evaluated on the basis of I-vector or probabilistic linear discriminant analysis (PLDA) scoring method. It is shown from SV experiments that a tandem feature of DNN bottleneck feature and MFCC feature gives the best performance when DNNs are configured using a rectangular type of hidden layers and trained with a supervised training method.