Smooth ROC curve estimation via Bernstein polynomials

The receiver operating characteristic (ROC) curve is commonly used to evaluate the accuracy of a diagnostic test for classifying observations into two groups. We propose two novel tuning parameters for estimating the ROC curve via Bernstein polynomial smoothing of the empirical ROC curve. The new estimator is very easy to implement with the naturally selected tuning parameter, as illustrated by analyzing both real and simulated data sets. Empirical performance is investigated through extensive simulation studies with a variety of scenarios where the two groups are both from a single family of distributions (symmetric or right skewed) or one from a symmetric and the other from a right skewed distribution. The new estimator is uniformly more efficient than the empirical ROC estimator, and very competitive to eleven other existing smooth ROC estimators in terms of mean integrated square errors.

Analysis of time distribution of passenger traffic on road transport on the basis of ride-sharing data

The paper studies indicators of passenger movement volumes on road transport for regional transport network that are gained on the basis of ride-sharing on-line services. For the processing of data arrays of common rides’ services web-scraping is used. The paper also investigates time distribution of passenger traffic that shows the filling of cars by passengers on various sections of a route. Peculiarities and dependencies of day fluctuations of passenger traffic for the Sverdlovsk region transport network are analysed. The author has considered a task of modeling of seasonal and cyclical fluctuations of passenger traffic volumes. She has used all necessary statistical procedures required for identification and assessment of model parameters and verification of its adequacy and accuracy. The time series model with polynomial smoothing is used for short-term forecasting that allows analysing the dynamics of passenger traffic, decreasing costs on monitoring of information on passenger traffic and allows using results for solving tasks of technological organization of public transport.

PROPERTIES OF DOUBLY ROBUST ESTIMATORS WHEN NUISANCE FUNCTIONS ARE ESTIMATED NONPARAMETRICALLY

An estimator of a finite-dimensional parameter is said to be doubly robust (DR) if it imposes parametric specifications on two unknown nuisance functions, but only requires that one of these two specifications is correct in order for the estimator to be consistent for the object of interest. In this article, we study versions of such estimators that use local polynomial smoothing for estimating the nuisance functions. We show that such semiparametric two-step (STS) versions of DR estimators have favorable theoretical and practical properties relative to other commonly used STS estimators. We also show that these gains are not generated by the DR property alone. Instead, it needs to be combined with an orthogonality condition on the estimation residuals from the nonparametric first stage, which we show to be satisfied in a wide range of models.

Comparison of standardised mortality ratios for renal failure before and after the 2011 Great East Japan Earthquake and Tsunami: an analysis of national vital statistics

ObjectiveThe impact of the 2011 Great East Japan Earthquake on renal failure (RF) risk remains unclear. We examined the 1-year impact of this disaster on RF mortality.SettingThis ecological study focused on the year before and after the earthquake. The data sources were national vital statistics (2010–2012), the national census (2010) and the Basic Resident Registration (2010–2012).ParticipantsOur study included all residents in Iwate, Miyagi and Fukushima, 1 year before and after the earthquake.Primary and secondary outcome measuresWe calculated standardised mortality ratios (SMRs) for RF, chronic RF and acute RF. Postearthquake weekly SMRs were calculated using the number of RF deaths for the corresponding weeks in 2010 as a reference. The SMRs for RF were compared between the coastal and inland municipalities using kernel-weighted polynomial smoothing.ResultsThere were 1290 RF deaths in the three prefectures during the year after the earthquake (chronic RF: 804 and acute RF: 236). The SMR for RF increased significantly in the first week after the earthquake in coastal areas (3.11; 95% CI: 1.84 to 4.37), but did not increase in inland areas (0.93; 95% CI: 0.47 to 1.38). A similar trend was observed for chronic RF (coastal: 4.0; 95% CI: 2.0 to 6.0; inland: 1.1; 95% CI: 0.4 to 1.7). SMRs for RF and chronic RF decreased over time and reached 1.0 approximately 20 weeks after the disaster. Changes in SMRs for acute RF were not apparent due to the low number of deaths.ConclusionsMortality due to RF and chronic RF, but not acute RF, increased in coastal areas after the earthquake. Chronic RF may have been exacerbated by disaster-induced sympathetic activation and poor management of renal dysfunction. Increased hypertension and damage to essential infrastructure and medical equipment may also have increased mortality in people with kidney disease.

Functional Coefficient Models for Economic and Financial Data

This article discusses the use of functional coefficient models for economic and financial data analysis. It first provides an overview of recent developments in the nonparametric estimation and testing of functional coefficient models, with particular emphasis on the kernel local polynomial smoothing method, before considering misspecification testing as an important econometric question when fitting a functional (varying) coefficient model or a trending time-varying coefficient model. It then describes two major real-life applications of functional coefficient models in economics and finance: the first deals with the use of functional coefficient instrumental-variable models to investigate the empirical relation between wages and education in a random sample of young Australian female workers from the 1985 wave of the Australian Longitudinal Survey, and the second is concerned with the use of functional coefficient beta models to analyze the common stock price of Microsoft stock (MSFT) during the year 2000 using the daily closing prices.