• The Korean Journal of Applied Statistics
• /
• v.30 no.5
• /
• pp.733-745
• /
• 2017

Local composite quantile regression is a useful non-parametric regression method widely used for its high efficiency. Data smoothing methods using kernel are typically used in the estimation process with performances that rely largely on the smoothing parameter rather than the kernel. However, $L_2$-norm is generally used as criterion to estimate the performance of the regression function. In addition, many studies have been conducted on the selection of smoothing parameters that minimize mean square error (MSE) or mean integrated square error (MISE). In this paper, we explored the optimality of selecting smoothing parameters that determine the performance of non-parametric regression models using local linear composite quantile regression. As evaluation criteria for the choice of smoothing parameter, we used mean absolute error (MAE) and mean integrated absolute error (MIAE), which have not been researched extensively due to mathematical difficulties. We proved the uniqueness of the optimal smoothing parameter based on MAE and MIAE. Furthermore, we compared the optimal smoothing parameter based on the proposed criteria (MAE and MIAE) with existing criteria (MSE and MISE). In this process, the properties of the proposed method were investigated through simulation studies in various situations.

• Journal of Biosystems Engineering
• /
• v.40 no.3
• /
• pp.296-304
• /
• 2015

Purpose: The balance between miscanthus production and its cost effectiveness depends greatly on its moisture content during post processing. The objective of this research was to measure the moisture content using a non-destructive and non-contact methodology for in situ applications. Methods: The moisture content of comminuted miscanthus was controlled using a closed chamber, a humidifier, a precision weigher, and a real-time monitoring software developed in this research. A CMOS sensor equipped with $50{\times}$ magnifier lens was used to capture magnified images of the conditioned materials with moisture content level from 5 to 30%. The hypothesis is that when light is incident on the comminuted particles in an inclined manner, higher moisture content results in light being reflected with a higher intensity. Results: A linear regression analysis for an initiative hypothesis based on general histogram analysis yielded insufficient correlations with low significance level (<0.31) for the determination coefficient. A significant relationship (94% confidence level) was determined at level 108 in a reverse accumulative histogram proposed based on a revised hypothesis. A linear regression model with the value at level 108 in the reverse accumulative histogram for a magnified image as the independent variable and the moisture content of comminuted miscanthus as the dependent variable was proposed as the estimation model. The calibrated linear regression model with a slope of 92.054 and an offset of 32.752 yielded 0.94 for the determination coefficient (RMSE = 0.2%). The validation test showed a significant relationship at the 74% confidence level with RMSE 6.4% (n = 36). Conclusions: To compensate the inconsistent significance between calibration and validation, an estimation model robust against various systematic interferences is necessary. The economic efficiency of miscanthus, which is a promising energy resource, can be improved by the real-time measurement of its crucial material properties.

• Korean Journal of Exercise Nutrition
• /
• v.25 no.1
• /
• pp.23-29
• /
• 2021

[Purpose] This preliminary study aimed to develop a regression model to estimate the non-exercise activity thermogenesis (NEAT) of Korean adults using various easy-to-measure dependent variables. [Methods] NEAT was measured in 71 healthy adults (male n = 29; female n = 42). Statistical analysis was performed to develop a NEAT estimation regression model using the stepwise regression method. [Results] We confirmed that age^A, weight^B, heart rate (HR)_average^C, weight × HR_average^D, weight × HR_sum^E, systolic blood pressure (SBP) × HR_rest^F, fat mass ÷ height^2G, gender × HR_average^H, and gender × weight × HR_sum^I were important variables in various NEAT activity regression models. There was no significant difference between the measured NEAT values obtained using a metabolic gas analyzer and the predicted NEAT. [Conclusion] This preliminary study developed a regression model to estimate the NEAT in healthy Korean adults. The regression model was as follows: sitting = 1.431 - 0.013 × (A) + 0.00014 × (D) - 0.00005 × (F) + 0.006 × (H); leg jiggling = 1.102 - 0.011 × (A) + 0.013 × (B) + 0.005 × (H); standing = 1.713 - 0.013 × (A) + 0.0000017 × (I); 4.5 km/h walking = 0.864 + 0.035 × (B) + 0.0000041 × (E); 6.0 km/h walking = 4.029 - 0.024 × (C) + 0.00071 × (D); climbing up 1 stair = 1.308 - 0.016 × (A) + 0.00035 × (D) - 0.000085 × (F) - 0.098 × (G); and climbing up 2 stairs = 1.442 - 0.023 × (A) - 0.000093 × (F) - 0.121 × (G) + 0.0000624 × (E).

• Communications for Statistical Applications and Methods
• /
• v.24 no.4
• /
• pp.383-396
• /
• 2017

The model in our approach assumes that computer responses are a realization of a Gaussian processes superimposed on a regression model called a Gaussian process regression model (GPRM). Selecting a subset of variables or building a good reduced model in classical regression is an important process to identify variables influential to responses and for further analysis such as prediction or classification. One reason to select some variables in the prediction aspect is to prevent the over-fitting or under-fitting to data. The same reasoning and approach can be applicable to GPRM. However, only a few works on the variable selection in GPRM were done. In this paper, we propose a new algorithm to build a good prediction model among some GPRMs. It is a post-work of the algorithm that includes the Welch method suggested by previous researchers. The proposed algorithms select some non-zero regression coefficients (${\beta}^{\prime}s$) using forward and backward methods along with the Lasso guided approach. During this process, the fixed were covariance parameters (${\theta}^{\prime}s$) that were pre-selected by the Welch algorithm. We illustrated the superiority of our proposed models over the Welch method and non-selection models using four test functions and one real data example. Future extensions are also discussed.

• The Journal of Asian Finance, Economics and Business
• /
• v.7 no.8
• /
• pp.161-173
• /
• 2020

This study examines the nonlinear relationship between financial development and economic growth in Pakistan using the threshold regression model for the period 1980-2017. We also employed quantile regression with 0.25, 0.50, and 0.75 quantiles of conditional distribution. The quantile regression is based on minimizing of sum of squared residuals. The result indicates that economic growth responds positively to financial development when the level of financial development surpasses the threshold value of 0.151. However, when financial development lies below the threshold value (that is, 0.151), its impact on economic growth is negative. Thus, when financial development of Pakistan surpasses the threshold level, it contributes more towards economic growth since greater level of financial development contributes more to boosts economic growth. This finding reveals that economic growth reacts differently to financial development, and the relationship between financial development and economic growth is U-shaped in Pakistan. Among the other variables, physical capital, labor force, and government expenditure exert a positive effect on economic growth. Furthermore, inflation rate and trade openness have an insignificant impact on economic growth. The results of quantile regression also confirm the non-linear relationship between financial development and economic growth in Pakistan. The finding of this study suggests revamping of financial sector policies in Pakistan.

• Horticultural Science & Technology
• /
• v.32 no.2
• /
• pp.241-251
• /
• 2014

This study was carried out to develop a simple, rapid and reliable assessment model to predict cold tolerance in Pittosporum tobira, a broad-leaved evergreen commonly used in the southern region of South Korea, which can minimize the possible experimental errors appeared in a electrolyte leakage test for cold tolerance assessment. The modeling procedure comprised of regrowth test and a electrolyte leakage test on the plants exposed to low temperature treatments. The lethal temperatures estimated from the methodological combinations of a electrolyte leakage test including tissue sampling, temperature treatment for potential electrical conductivity, and statistical analysis were compared to the results of the regrowth test. The highest temperature showing the survival rate lower than 50% obtained from the regrowth test was $-10^{\circ}C$ and the lethal was $-10^{\circ}C{\sim}-5^{\circ}C$. Based on the results of the regrowth test, several methodological combinations of electrolyte leakage tests were evaluated and the electrolyte leakage lethal temperatures estimated using leaf sample tissue and freeze-killing method were closest to the regrowth lethal temperature. Evaluating statistical analysis models, linear interpolation had a higher tendency to overestimate the cold tolerance than non-linear regression. Consequently, the optimal model for cold tolerance assessment of P. tobira is composed of evaluating electrolyte leakage from leaf sample tissue applying freeze-killing method for potential electrical conductivity and predicting lethal temperature through non-linear regression analysis.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.33 no.5
• /
• pp.1809-1827
• /
• 2013

Recently we need to take various measures to prepare for extreme flood that occur due to climate change. It is important that establish flood forecasting system to prepare flood over non-structure measures. The objective of this study is to develop superior real-time flood forecasting model by comparing the Neuro-fuzzy model and the multiple linear regression model. The Neuro-fuzzy model and the multiple linear regression model are established using same input data and applied for various flood events in Nakdong basin. The results show that the Neuro-fuzzy model can carry out flood forecasting results more accurately than the multiple linear regression model. This study can contribute to the establishment of a high accuracy flood information system that secure lead time in Nakdong basin.

• Proceedings of the Korean Society of Near Infrared Spectroscopy Conference
• /
• 2001.06a
• /
• pp.1524-1524
• /
• 2001

A previous paper(Ito et al., 2000) has described the improvement of the standard error(SEC and SEP) of the predicted soluble solids(Brix) in a melon cultivar by non-contact mode with a fiber optic probe. Then we examined the immature and mature fruits. The objective of this study was to determine if non-contact mode could improve the standard error of the predicted Brix of matured melon fruits from cross progeny as well as the contact mode(usual method). The optical absorption spectrum was measured using a NIR Systems model 6500 spectrophotometer. A commercial spectral program(NSAS ver. 3.27) was used for multiple linear regression analysis. Absorbances of 902 and in the vicinity of 877 nm were included as the independent variables in both multiple regression equations. These wavelengths are key wavelengths for non-destructive Brix determination. When the results for the contact mode and non-contact mode are compared, the latter mode improved the former standard error(SEP and RMS).

• Journal of Wetlands Research
• /
• v.14 no.3
• /
• pp.341-352
• /
• 2012

The purpose of this study is to statistically project future probable rainfall and to quantitatively assess a future flood vulnerability using flood vulnerability model. To project probable rainfall under non-stationarity conditions, the parameters of General Extreme Value (GEV) distribution were estimated using the 1 yr data added to the initial 30 yr base series. We can also fit a linear regression model between time and location parameters after comparing the linear relationships between time and location, scale, and shape parameters, the probable rainfall in 2030 yr was calculated using the location parameters obtained from linear regression equation. The flood vulnerability in 2030 yr was assessed inputted the probable rainfall into flood vulnerability assessment model suggested by Jang and Kim (2009). As the result of analysis, when a 100 yr rainfall frequency occurs in 2030 yr, it was projected that vulnerability will be increased by spatial average 5 % relative to present.

• Bulletin of the Korean Chemical Society
• /
• v.18 no.9
• /
• pp.1006-1010
• /
• 1997

The nature and analysis methods of complicated decay profiles found in forced Rayleigh scattering (FRS) have been investigated for the probe diffusion of methyl yellow in 2-propanol. The complementary shifted and ground state grating effect, which is known to be the origin of non-single exponential decays, was analyzed by non-linear regression fitting to a double exponential model function. We confirmed that the parameters were highly correlated so that it was difficult to extract a unique set of parameters in the presence of experimental noise. Nevertheless, a reasonable range of decay time constants could be estimated from the grating spacing dependence.