Parametric Regression Techniques Research Articles

Explaining Violent Intra-Ethnic Conflict

Despite significant advances in the disaggregation of the study of civil conflict and intra-ethnic violence, intra-ethnic violence remains understudied. In this article, we present the first systematic, cross-national analysis of the conditions that promote violent, fragmentary conflict within politically active ethnic minorities. We propose a model of intra-ethnic conflict in which collective violence is produced by the interaction between subgroup entrepreneurs and the suppressive actions of the state. This two-level model predicts a curvilinear relationship between the relative size of an ethnic minority and its probability of experiencing large-scale intra-ethnic conflict. Additional hypotheses based on the proposed causal mechanism are also posited. These hypotheses are tested with data drawn from a global sample of politically active ethnic minorities, for the period 1990 through 2006, using a combination of parametric and semi-parametric regression techniques. The results strongly confirm the predicted curvilinear relationship while also demonstrating that the specific shape of this relationship shifts in predictable ways under varying social and political contexts.

Probabilistic Forecasting

A probabilistic forecast takes the form of a predictive probability distribution over future quantities or events of interest. Probabilistic forecasting aims to maximize the sharpness of the predictive distributions, subject to calibration, on the basis of the available information set. We formalize and study notions of calibration in a prediction space setting. In practice, probabilistic calibration can be checked by examining probability integral transform (PIT) histograms. Proper scoring rules such as the logarithmic score and the continuous ranked probability score serve to assess calibration and sharpness simultaneously. As a special case, consistent scoring functions provide decision-theoretically coherent tools for evaluating point forecasts. We emphasize methodological links to parametric and nonparametric distributional regression techniques, which attempt to model and to estimate conditional distribution functions; we use the context of statistically postprocessed ensemble forecasts in numerical weather prediction as an example. Throughout, we illustrate concepts and methodologies in data examples.

COST-EFFECTIVENESS OF HUMAN LEUKOCYTE ANTIGEN MATCHING IN PENETRATING KERATOPLASTY.

Background: The matching of favorable human leukocyte antigen (HLA) combinations is rarely performed in penetrating keratoplasty procedures for primary prophylaxis of immune reactions. However, clinical studies suggest that the incidence of graft rejection decreases substantially when patients receive favorably matched grafts. Objective: The aim of this study was to assess the cost-effectiveness of HLA matching for patients undergoing penetrating keratoplasty in everyday clinical practice. Methods: In the absence of a randomized controlled clinical trial, we used administrative data from the Freiburg University Eye Hospital in Germany. Our study population consisted of all patients who underwent their first keratoplasty between 11/2003 and 01/2010 and for whom information on HLA histocompatibility was available. We used propensity score matching to estimate a causal effect of favorable HLA matching, parametric survival regression techniques to predict graft survival and expert opinion to model incremental cost for HLA matching. Because the availability of favorable HLA histocompatibility ultimately depends on the patients' HLA phenotype, we modeled the incremental cost-effectiveness ratio (ICER) as a function of the probability that a patient will receive a favorably matched HLA, and used expert opinion to set a point estimate. Results: We predicted that corneal grafts with favorable HLA matching were associated with improved rejection-free graft survival time (more than 1,000 days). We estimated the incremental cost of HLA matching at EUR 1,200 and the ICER at EUR 4.62 per additional day of graft survival. Conclusions: The ICER of HLA matching is acceptable, given the high cost of alternative treatment and the shortage of corneal donors in Germany.

Suitability of KNN Regression in the Development of Interaction based Software Fault Prediction Models

Accurate fault prediction is an indispensable step, to the extent of being a critical activity in software engineering. In fault prediction model development research, combination of metrics significantly improves the prediction capability of the model, but it also gives rise to the issue of handling an increased number of predictors and evolved nonlinearity due to complex interaction among metrics.Ordinary least square (OLS) based parametric regression techniques cannot effectively model such nonlinearity with a large number of predictors because the global parametric function to fit the data is not known beforehand. In our previous studies[1–3], we showed the impact of interaction in the combined metrics approach of fault prediction and statistically established the simultaneous increment in the predictive accuracy of the model with interaction.In this study we use K-Nearest Neighbor (KNN) regression as an example of nonparametric regression technique, otherwise well known for classification tasks in the data mining community. Through the results derived here, we empirically establish and validate the hypothesis that the performance of KNN regression remains ordinarily unaffected with increasing number of interacting predictors and simultaneously provides superior performance over widely used multiple linear regression (MLR).

Vertical Spillover Effects of Multinationals on Chinese Domestic Firms via Supplier–Customer Relationships

AbstractForeign direct investment (FDI) can benefit domestic firms in the host country. Using firm‐level data for China, we find statistically positive vertical spillover effects of multinational enterprises on the performance of domestic firms through backward and forward supplier–customer relationships. The spillover effects are mainly from large multinational enterprises and are greater for state‐owned firms and in poor regions. Our results are robust for both parametric regression and nonparametric matching techniques. Our findings have strong policy implications: while regulations relating to building business relationships with domestic firms when seeking foreign direct investment should be established, such policies should be aimed at private firms, big multinationals and less developed regions.

Is Posner Right? An Empirical Test of the Posner Argument for Transferring Health Spending from Old Women to Old Men

Posner (Aging and old age, University of Chicago Press, Chicago, 1995) proposes the redistribution of health spending from old women to old men to equalize life expectancy. His argument is based on the assumption that the woman’s utility is higher if her husband is alive. Using self-reported satisfaction measures from a long-running German panel survey, the Socio-Economic Panel Study (SOEP), the present study conducts an empirical test of this assumption and investigates the question of whether and to what extent widowed women’s utility responds to her spouse’s death. We apply a combination of propensity score matching and parametric regression techniques. Our results reveal satisfaction trajectories of women who experience the death of their spouse and identifies the causal effect of widowhood. The average level of satisfaction in a control group of non-widowed women serves as a reference to measure the degree of adaptation to widowhood. The results suggest bereavement has no enduring effect on satisfaction, and that is evidence against Posner’s assumption. We conclude that elderly women would not benefit from Posners policy proposal.

Regional flood frequency analysis in eastern Australia: Bayesian GLS regression-based methods within fixed region and ROI framework – Quantile Regression vs. Parameter Regression Technique

Summary In this article, an approach using Bayesian Generalised Least Squares (BGLS) regression in a region-of-influence (ROI) framework is proposed for regional flood frequency analysis (RFFA) for ungauged catchments. Using the data from 399 catchments in eastern Australia, the BGLS-ROI is constructed to regionalise the flood quantiles (Quantile Regression Technique (QRT)) and the first three moments of the log-Pearson type 3 (LP3) distribution (Parameter Regression Technique (PRT)). This scheme firstly develops a fixed region model to select the best set of predictor variables for use in the subsequent regression analyses using an approach that minimises the model error variance while also satisfying a number of statistical selection criteria. The identified optimal regression equation is then used in the ROI experiment where the ROI is chosen for a site in question as the region that minimises the predictive uncertainty. To evaluate the overall performances of the quantiles estimated by the QRT and PRT, a one-at-a-time cross-validation procedure is applied. Results of the proposed method indicate that both the QRT and PRT in a BGLS-ROI framework lead to more accurate and reliable estimates of flood quantiles and moments of the LP3 distribution when compared to a fixed region approach. Also the BGLS-ROI can deal reasonably well with the heterogeneity in Australian catchments as evidenced by the regression diagnostics. Based on the evaluation statistics it was found that both BGLS-QRT and PRT-ROI perform similarly well, which suggests that the PRT is a viable alternative to QRT in RFFA. The RFFA methods developed in this paper is based on the database available in eastern Australia. It is expected that availability of a more comprehensive database (in terms of both quality and quantity) will further improve the predictive performance of both the fixed and ROI based RFFA methods presented in this study, which however needs to be investigated in future when such a database is available.

Trends in extreme events of rainfall in low country wet zone of Sri Lanka

It has been reported that growth and production of crops are severally affected by the changing climate. Thus, climate changes may have significant impact on a country like Sri Lanka which has an agriculture based economy. Plantation sector contributes more to export agriculture, which is mainly confined in the wet zone of the Island. Most of the management practices of plantation crops such as tea, rubber and coconut are linked with weather pattern; especially with the rainfall pattern. Evidence for change in the amount and the distribution of rainfall in Sri Lanka are not very apparent. However, trends in the occurrence of extreme events of rainfall may be apparent and there by the crop production may be affected. A study was carried out to identify trends in extreme events of rainfall in low country wet zone where most of rubber plantations and low grown tea plantations are found. Daily rainfall data from eight agro ecological stations in the Wet zone Low country (WL) of the Island for the period from 1941 to 2010 were used in the study. Boundary line techniques, parametric and non parametric regression techniques were used to identify trends in extreme events of rainfall. It was found that significant changes in total precipitation were not apparent while there were very clear evidences for changing incidences of extreme events with respect to rainfall. In some areas under investigation, trends of occurring low annual rainfalls in time was apparent. There were significant evidences for declining trends in annual rainy days in some areas of low country wet zone of the country, which may be favorable for operations like exploitation of rubber latex. However, there was a tendency for increasing incidences of lengthy dry spells in some areas in the low country wet zone. This may bring adverse impacts on some operations such as crop establishment and fertilizer application and crop yield of tea and rubber.

Regional flood frequency analysis using Bayesian generalized least squares: a comparison between quantile and parameter regression techniques

AbstractRegression‐based regional flood frequency analysis (RFFA) methods are widely adopted in hydrology. This paper compares two regression‐based RFFA methods using a Bayesian generalized least squares (GLS) modelling framework; the two are quantile regression technique (QRT) and parameter regression technique (PRT). In this study, the QRT focuses on the development of prediction equations for a flood quantile in the range of 2 to 100 years average recurrence intervals (ARI), while the PRT develops prediction equations for the first three moments of the log Pearson Type 3 (LP3) distribution, which are the mean, standard deviation and skew of the logarithms of the annual maximum flows; these regional parameters are then used to fit the LP3 distribution to estimate the desired flood quantiles at a given site. It has been shown that using a method similar to stepwise regression and by employing a number of statistics such as the model error variance, average variance of prediction, Bayesian information criterion and Akaike information criterion, the best set of explanatory variables in the GLS regression can be identified. In this study, a range of statistics and diagnostic plots have been adopted to evaluate the regression models. The method has been applied to 53 catchments in Tasmania, Australia. It has been found that catchment area and design rainfall intensity are the most important explanatory variables in predicting flood quantiles using the QRT. For the PRT, a total of four explanatory variables were adopted for predicting the mean, standard deviation and skew. The developed regression models satisfy the underlying model assumptions quite well; of importance, no outlier sites are detected in the plots of the regression diagnostics of the adopted regression equations. Based on ‘one‐at‐a‐time cross validation’ and a number of evaluation statistics, it has been found that for Tasmania the QRT provides more accurate flood quantile estimates for the higher ARIs while the PRT provides relatively better estimates for the smaller ARIs. The RFFA techniques presented here can easily be adapted to other Australian states and countries to derive more accurate regional flood predictions. Copyright © 2011 John Wiley & Sons, Ltd.

Influence of climate, edaphic factors and tree nutrition on site index of chestnut coppice stands in north-west Spain

The relationships between climatic variables, soil parameters, tree nutrition and site index (SI) were examined in 30 chestnut coppice stands in Asturias, the region with the largest area of chestnut coppice in north-west Spain. SI was estimated from a top height projection equation for this species. A soil sample was obtained at 0- to 20-cm depth and several leaves were collected from the upper zone of the dominant trees in each plot for further analysis. Chi-square automatic interaction detection and parametric regression techniques were used to determine the key factors affecting SI. According to the regression tree and the parametric regression model for all variables, the extractable Mg in the soil and mean temperature of the warmest month were the most important variables explaining SI in the region. Elevation and foliar concentrations of phosphorus were not significant factors in the parametric regression analysis (although considered key factors when foliar nutrients and physiographical variables were analysed separately). The results show the importance, for chestnut growth, of site selection and balanced fertilizer treatment consistent with the soil requirements of the species.

Effects of foliar nutrients and environmental factors on site productivity in Pinus pinaster Ait. stands in Asturias (NW Spain)

The effects of foliar nutrients and environmental variables on site productivity in Maritime pine (Pinus pinaster Ait.) stands in Asturias (NW Spain) were studied. Site productivity was evaluated by means of site index. The environmental characteristics considered included physiographic, edaphic and climatic variables. Two different statistical analyses were carried out to determine the key factors affecting site productivity: the CHAID procedure and parametric regression techniques. The regression trees obtained by applying the CHAID algorithm separately to each type of variable revealed the order of importance of the environmental factors. In addition, the variables corresponding to the first level of branches of the CHAID regression trees were consistent with those identified by parametric regression. According to the parametric regression model, a physical soil characteristic (soil depth) and a climatic characteristic (mean summer temperature) were the most important variables for explaining site productivity in the region. Foliar concentrations of phosphorus and elevation were not significant factors in the parametric regression analysis (although considered key factors when foliar nutrients and physiographic variables were analyzed separately). Although the percentage variability in SI explained by the best parametric model was not very high (45%), the model is parsimonious and provides robust estimates of the spatial distribution of this variable. The results obtained provide further information about the ecology of this species in Asturias, and represent an important tool in forest management, especially for afforested areas.

Monsoon rainfall extreme indices and tendencies from 1954–2003 in Kerala, India

Climate change has the potential ability to alter the occurrence and severity of extreme events. Though predicting changes of such extreme events is difficult, understanding them is important to determine the impacts of climate change in various sectors. This paper presents the change in rainfall extremes in the monsoon season in south-west Indian peninsula. Daily rainfall data were analysed for the entire Kerala state in India to determine if the extreme rainfall had changed over the 50-year period. Several indices were derived from the data to identify the extreme rainfalls. The trends of all the extreme indices were assessed by parametric ordinary least square regression technique, which were tested for significance at 95% level. Results showed significant decrease in monsoon rainfall extremes in Kerala that would affect the tendency of change in seasonal total rainfall. This study provides a comprehensive knowledge on extreme monsoon precipitation in Kerala, which could also be employed to study changing climate at local scale in other regions.

Effect of Age and Pathologic Gleason Score on PSA Recurrence: Analysis of 2911 Patients Undergoing Radical Prostatectomy

To clarify the relationship between age and pathologic Gleason score and their effect on prostate-specific antigen recurrence (PSAR). The data from a cohort of 2911 men who had undergone radical prostatectomy from 1988 to 2006 were retrieved from the Duke Prostate Center database. Patient age was divided into 3 groups: <60, 60-64, and >or=65 years. The pathologic Gleason score was divided into 5 groups: <or=5, 6, 3 + 4, 4 + 3, and >7. PSAR was defined as the prostate-specific antigen level increasing to >0.2 ng/mL >30 days after radical prostatectomy. The associations between age and pathologic Gleason score on PSAR and the time to PSAR were analyzed using parametric, nonparametric, Kaplan-Meier, and Cox regression techniques. Patient age and interval to PSAR had no significant association (P > .05). Kaplan-Meier analysis demonstrated a significant difference in PSAR among age groups. The pathologic Gleason scores of 3 + 3, 3 + 4, 4 + 3, and >7 were significant in determining the incidence of PSAR. Age was not significant for PSAR in patients with a pathologic Gleason score of <or=7. In patients with a pathologic Gleason score of >7, a statistically significant difference was observed among the age groups. Men <60 years old with a pathologic Gleason score >7 had a lower incidence of PSAR than did older men with a similar pathologic Gleason score. A pathologic Gleason score of >or=6 was significant in predicting PSAR. Age alone was an independent factor in predicting PSAR, but not in predicting the interval to PSAR. The pathologic Gleason score remained a predictor of PSAR, and patient age should be considered in patients with a pathologic Gleason score >7.

Long-term changes and variability of monthly extreme temperatures in India

Extreme temperatures are changing worldwide together with changes in the mean temperatures. This study investigates the long-term trends and variations of the monthly maximum and minimum temperatures and their effects on seasonal fluctuations in various climatological regions in India. The magnitude of the trends and their statistical significance were determined by parametric ordinary least square regression techniques and the variations were determined by the respective coefficient of variations. The results showed that the monthly maximum temperature increased, though unevenly, over the last century. Minimum temperature changes were more variable than maximum temperature changes, both temporally and spatially, with results of lesser significance. The results of this study are good indicators of Indian climate variability and its changes over the last century.

Parametric Robust Regression Analysis of Contaminated Data

Summary In reality it is more often the case when part of data have been drawn from some distribution other than that we believe they have been drawn from. Traditional nonparametric or parametric statistics are insufficient for such contaminated data. Tsou [Tsou, T.-S. (2003). Parametric robust inferences for regression parameters under generalized linear models. Submitted] proposed parametric robust regression techniques that provide asymptotically valid inference for regression parameters so long as the true distributions that generate the data have finite second moments. It will be demonstrated that the novel robust regression tools fit well for the regression analysis of data generated from unknown and distinct families of distributions.

Structural Changes And Models Of Inflation In South Africa

In the past three decades South Africa has undergone a number of structural changes in the economy, which may have contributed to parameter instability in inflation equations. In this study the out-of-sample forecasting performance of three single equation models for inflation in South Africa is examined without imposing the restriction that coefficients are fixed over time. A time-varying parameter regression (VPR) technique based on the recursive application of the Kalman filter is used to evaluate the stability of the models using South African quarterly data from 1971.1 to 1998.3, with 1998.4 to 2001.3 excluded from the estimated sample period in order to do an ex post forecast evaluation. It is found that the parameters of the expectations-augmented Phillips curve and the traditional monetarist models for inflation are fairly stable over the sample period, whereas the money demand model for inflation shows signs of changing parameters over time. This is in accordance with some changes in monetary policy during the 1980s and a degree of financial liberalisation since 1994. It is furthermore shown that, under conditions of structural instability, the VPR money demand model can substantially reduce out-of-sample forecasting errors compared to its fixed-coefficient counterparts.

Business groups: entry barrier–innovation debate revisited

We argue that in the presence of market imperfections, business groups’ market share in a sector acts as a proxy for barriers to enter that sector. Recognition of group share as a proxy for entry barriers allows us to revisit the entry barrier–innovation debate in the context of emerging economies. Evidence presented in this paper note innovative performance is at its peak when groups’ market share is at an intermediate level, suggesting that either too many or too few entry barriers could be harmful for innovation. This result appears robust across a variety of non-parametric and parametric regression techniques.

Continuous age-dependent reference ranges for thyroid hormones in neonates, infants, children and adolescents established using the ADVIA Centaur Analyzer.

In neonates, infants, children and adolescents serum concentrations of the thyroid hormones show a clear age dependency. Currently, age-dependent thyroid hormone reference ranges for the new ADVIA Centaur immunoanalyzer are not available. Thyroid-stimulating hormone (TSH), free thyroxine (FT4), total thyroxine (T4), free triiodothyronine (FT3) and total triiodothyronine (T3) were determined in the sera of 460 healthy children aged 0 daysto 18 years from an inland (n=308) and a seaside city (n=152) using the Bayer ADVIA Centaur analyzer. Square root functions defining the upper and lower limit of the continuous age-dependent reference range and the median line for each thyroid hormone were estimated applying a parametric regression technique. Continuous age-dependent reference ranges show a better fit to the obtained data than discontinuous reference ranges. The median TSH and FT4 concentrations decrease within the first year of life and are nearly constant until 18 years of age. The median T4 decreases within the entire age interval, whereas the median FF3 is constant. The median T3 increases within the first year of life and decreases afterwards. The calculated continuous reference ranges for T4, T3 and partly FT4 are consistent with published reference ranges. However, the reference ranges obtained for FT3 and partly TSH differed considerably, compared to published reference ranges. No significant residence- or sex-specific effects on age-adjusted hormone concentrations were found except for a slight residence-specific effect on age-adjusted FT4, which was probably due to assay imprecision. Due to the lack of international standardization, the assay-specific evaluation of reference ranges is very important. Continuous age-dependent reference ranges comply better with the biological reality and are more reliable than discontinuous reference ranges.

Least Committed Splines in 3D Modelling of Free Form Objects from Intensity Images

Generating 3D models of objects from video sequences is an important problem in many multimedia applications ranging from teleconferencing to virtual reality. In this paper, we present a method of estimating the 3D face model from a monocular image sequence, using a few standard results from the affine camera geometry literature in computer vision, and spline fitting techniques using a modified non parametric regression technique. We use the bicubic spline functions to model the depth map, given a set of observation depth maps computed from frame pairs in a video sequence. The minimal number of splines are chosen on the basis of the Schwartz's Criterion. We extend the spline fitting algorithm to hierarchical splines. Note that the camera calibration parameters and the prior knowledge of the object shape is not required by the algorithm. The system has been successfully demonstrated to extract 3D face structure of humans as well as other objects, starting from their image sequences.

Nonparametric identification of daily activity durations using kernel density estimators

Modeling of activity duration has become an important aspect in characterizing activity and travel patterns. The standard approach for analyzing activity duration is to use hazard-based models to account for the duration dependence within an activity while estimating covariate effects and heterogeneity. The effects of covariates and heterogeneity on duration have been modeled using parametric and nonparametric regression techniques. In this paper, we present a nonparametric pattern recognition approach that can precede hazard-based duration modeling to identify heterogeneity patterns that may undermine duration modeling if not detected. The technique utilizes a kernel estimate of the probability density function (pdf) of various activity durations and allows for statistical comparison of distributions to evaluate differences between groups of individuals. Preliminary testing on the first wave of a panel survey indicates that the approach is insightful in evaluating covariate effects while allowing visual inspection of heterogeneity in the data.