The Impact of Foreign Direct Investment on the Economic Growth of South Caucasus Countries

This paper focuses on the effects of foreign direct investment inflows on the economic growth in a panel of three South Caucasus countries using data from 1996-2019 periods. In this study, we applied the following control variables; trade openness, investment, real exchange rate, and population growth. Classical linear regression model was employed in this paper. Ordinary least squares methods are used for estimation. Empirical results revealed that there is no significant effect of FDI inflows on economic growth. The results show that inward FDI stock-to-GDP ratio and real GDP growth rate are positively correlated.

Subsampling and Aggregation: A Solution to the Scalability Problem in Distance-Based Prediction for Mixed-Type Data

The distance-based linear model (DB-LM) extends the classical linear regression to the framework of mixed-type predictors or when the only available information is a distance matrix between regressors (as it sometimes happens with big data). The main drawback of these DB methods is their computational cost, particularly due to the eigendecomposition of the Gram matrix. In this context, ensemble regression techniques provide a useful alternative to fitting the model to the whole sample. This work analyzes the performance of three subsampling and aggregation techniques in DB regression on two specific large, real datasets. We also analyze, via simulations, the performance of bagging and DB logistic regression in the classification problem with mixed-type features and large sample sizes.

Diel Variation of Viral Production in a Coastal Subtropical Marine System

Viral production (VP) and bacterial mortality by viral lysis critically influence the production and mortality of aquatic bacteria. Although bacterial production, mortality by viral lysis, and viral density have been found to exhibit diel variations, the diel change in viral production has rarely been investigated. In this study, we conducted two diel dilution incubation experiments in a semi-enclosed, nutrient-rich coastal region in northeastern Taiwan to estimate the diel viral production and the mortality by viral lysis. We also compared two methods (linear regression between viral density and time versus arithmetic mean of VP during incubation) of estimating viral production. We found that viral production estimated by linear regression and bacterial mortality by viral lysis were higher during the daytime than during the nighttime. A possible explanation for the high viral production at daytime is that the bacterial community was composed of cell types with higher burst sizes at daytime. We further argued that the classical linear regression method can be used only when viral density significantly linearly increases with time, which does not always occur in dilution incubations. This study offered observations of diel variation in viral dynamics and discussed the methods estimating viral production in a marine environment.

UNIFORM-IN-SUBMODEL BOUNDS FOR LINEAR REGRESSION IN A MODEL-FREE FRAMEWORK

For the last two decades, high-dimensional data and methods have proliferated throughout the literature. Yet, the classical technique of linear regression has not lost its usefulness in applications. In fact, many high-dimensional estimation techniques can be seen as variable selection that leads to a smaller set of variables (a “submodel”) where classical linear regression applies. We analyze linear regression estimators resulting from model selection by proving estimation error and linear representation bounds uniformly over sets of submodels. Based on deterministic inequalities, our results provide “good” rates when applied to both independent and dependent data. These results are useful in meaningfully interpreting the linear regression estimator obtained after exploring and reducing the variables and also in justifying post-model-selection inference. All results are derived under no model assumptions and are nonasymptotic in nature.

Physical Activity and Cumulative Long-Term Care Cost among Older Japanese Adults: A Prospective Study in JAGES

This study aimed to determine the impact of physical activity on the cumulative cost of long-term care insurance (LTCI) services in a cohort of community-dwelling people (65 years and older) in Japan. Using cohort data from the Japan Gerontological Evaluation Study (JAGES) on those who were functionally independent as of 2010/11, we examined differences in the cumulative cost of LTCI services by physical activity. We followed 38,875 participants with LTCI service costs for 59 months. Physical activity was assessed by the frequency of going out and time spent walking. We adopted a generalized linear model with gamma distribution and log-link function, and a classical linear regression with multiple imputation. The cumulative LTCI costs significantly decreased with the frequency of going out and the time spent walking after adjustment for baseline covariates. LTCI’s cumulative cost for those who went out once a week or less was USD 600 higher than those who went out almost daily. Furthermore, costs for those who walked for less than 30 min were USD 900 higher than those who walked for more than 60 min. Physical activity among older individuals can reduce LTCI costs, which could provide a rationale for expenditure intervention programs that promote physical activity.

A Statistical Approach to Evaluating Wheel-Rail Contact Dynamics

This paper provides a statistical analysis of the effects of wheel load, angle of attack (AoA), and creepage on longitudinal traction force at the wheel-rail contact using experimental data collected on the Virginia Tech-Federal Railroad Administration (VT-FRA) Roller Rig. The VT-FRA Roller Rig is a unique piece of equipment designed and built with the specific goal of evaluating the wheel-rail contact mechanics and dynamics with a high degree of precision. Longitudinal traction forces are of great importance to the railroad industry since they provide the motive power needed to move a train. Various experiments are conducted in different settings to study the relationship between the aforementioned variables and the longitudinal traction force. The test data is split into “training” and “testing” sets, and the training sets (a total of four) are used to entertain statistical models in a standard parametric regression framework. The study carefully assesses whether the assumptions of the classical linear regression model hold by studying the empirical histogram and the normal Q-Q plot of the residuals. In the case of non-linearities, different transformations are applied to the explanatory variable to find the closest functional form that captures the relationship between the response and the explanatory variables. The developed models are then compared with their non-parametric counterparts such as natural cubic splines in terms of goodness of fit, and prediction error on the testing set. The study develops regression models that are able to accurately explain the relationship between longitudinal traction and creepage and AoA. The models are intended to be used for predicting traction under various operating conditions.

Towards Improving Transparency of Count Data Regression Models for Health Impacts of Air Pollution

In studies on the health impacts of air pollution, regression analysis continues to advance far beyond classical linear regression, which many scientists may have become familiar with in an introductory statistics course. With each new level of complexity, regression analysis may become less transparent, even to the analyst working with the data. This may be especially true in count data regression models, where the response variable (typically given the symbol y) is count data (i.e., takes on values of 0, 1, 2, …). In such models, the normal distribution (the familiar bell-shaped curve) for the residuals (i.e., the differences between the observed values and the values predicted by the regression model) no longer applies. Unless care is taken to correctly specify just how those residuals are distributed, the tendency to accept untrue hypotheses may be greatly increased. The aim of this paper is to present a simple histogram of predicted and observed count values (POCH), which, while rarely found in the environmental literature but presented in authoritative statistical texts, can dramatically reduce the risk of accepting untrue hypotheses. POCH can also increase the transparency of count data regression models to analysts themselves and to the scientific community in general.

Violation of the assumption of homoscedasticity and detection of heteroscedasticity

In this paper, it is assumed that there is a violation of homoskedasticity in a certain classical linear regression model, and we have checked this with certain methods. Model refers to the dependence of savings on income. Proof of the hypothesis was performed by data simulation. The aim of this paper is to develop a methodology for testing a certain model for the presence of heteroskedasticity. We used the graphical method in combination with 4 tests (Goldfeld-Quantum, Glejser, White and Breusch-Pagan). The methodology that was used in this paper showed that the assumption of homoskedasticity was violated and it showed existence of heteroskedasticity.

NON-HYBRID ENSEMBLE SPATIAL REGRESSION ON HUMAN DEVELOPMENT INDEX (IPM) in CENTRAL JAVA

The human development index (HDI) is a measure to see an increase in regional development that has a very broad dimension, because it increases the quality of the population of an area in terms of life expectancy, education, and decent standard of living. In 2010 the Central Java HDI increased by 66.08% and increased by 4.44%, with the total HDI in 2017 of 70.52 percent. Spatial regression is the development of classical linear regression involving the region model. Spatial regression ensemble is a technique to be sent spasi spatial regression models by adding noise (additive noise). The type of spatial weighting used is Queen Contiguity. The selection of the best model using AIC and RMSE values. The purpose of this study is to provide an assessment of the distribution of HDI data in the Province of Central Java in 2017 and to do modeling using non-hybrid spatial ensemble regression regression. The results of this study are the SAR spatial method with ensemble giving results with AIC value of 143 and RMSE value of 1.3899 with a value of 90.09%. Significant variables on HDI are population density (X1), poverty (X2), school participation rates (X5), and average per capita per month for food and non-food (X7).