Lianfen Qian

Interests

Research

My research interests lie at the intersection of theoretical and applied statistics, with an emphasis on developing rigorous statistical methodology motivated by real-world data problems. I am particularly interested in survival analysis and event-history modeling, including frailty models, semi-competing risks, cure models, and change-point detection under censoring and measurement error. A central theme of my work is asymptotic parametric and nonparametric inference, weak convergence of stochastic processes, and their applications to longitudinal data, time series, and high-dimensional settings. I also have strong interests in data analytics, statistical machine learning, and financial mathematics, including modeling financial time series, risk management, and option trading strategies. More recently, my research has expanded to interdisciplinary applications in biomedical and environmental sciences, consumer finance, and the responsible use of AI tools in higher education. Across these areas, my goal is to bridge theory and practice by developing statistically sound, computationally efficient methods that support robust inference and informed decision-making.

Teaching

My teaching interests span theoretical and applied statistics, actuarial science, data science, and financial mathematics, with a particular emphasis on integrating rigorous mathematical foundations with modern computational tools. I am especially interested in teaching mathematical statistics, regression and generalized linear models, statistical machine learning, survival analysis, financial mathematics, and time series analysis. Given my experience teaching both undergraduate and graduate courses across mathematics, statistics, business analytics, and biostatistics, I am committed to designing curricula that bridge theory and practice while preparing students for careers in quantitative finance, data science, research, and industry. I am also interested in advancing AI-enhanced pedagogy, reproducible research training, and experiential learning models that incorporate real-world datasets, interdisciplinary collaboration, and industry-relevant problem solving.

Courses

Scholarly Contributions

Journals/Publications

• Feng, W., Spohn, D., Qian, L., & Hassan, M. K. (2025). Consumer financial anxiety during the COVID-19 pandemic. Borsa Istanbul Review. doi:10.1016/j.bir.2025.08.008
  More info

  This paper examines the determinants of consumer financial anxiety using data from the 2021 National Financial Capability Study (NFCS), which covers the COVID-19 pandemic. Using an ordinal logistic regression, we control for demographic variables and demonstrate that negative events, such as job loss and income reduction, significantly increase financial anxiety, whereas having precautionary savings substantially reduces it. Then, we use a partial least squares structural equation model (PLS-SEM) to study the impacts of financial literacy and financial practices on anxiety. Our findings reveal that, whereas financial literacy has a modest direct effect, positive financial behaviors, such as saving and budgeting, play a significantly greater role in alleviating financial anxiety. The empirical lessons from the study and the analytical framework that we propose extend to post-pandemic consumer financial well-being, with important implications for policy, education, and mental health interventions.
• Zhang, C., Xu, K., & Qian, L. (2020). Asymptotic properties of the QMLE in a log-linear RealGARCH model with Gaussian errors. Statistical Papers, 61(Issue 6). doi:10.1007/s00362-018-1051-8
  More info

  To incorporate the realized volatility in stock return, Hansen et al. (J Appl Econ 27:877–906, 2012) proposed a RealGARCH model and conjectured some theoretical properties about the quasi-maximum likelihood estimation (QMLE) for parameters in a log-linear RealGARCH model without rigorous proof. Under Gaussian errors, this paper derives the detailed proof of the theoretical results including consistency and asymptotic normality of the QMLE, hence it solves the conjectures in Hansen et al. (J Appl Econ 27:877–906, 2012).
• Schesser Bartra, S., Lorica, C., Qian, L., Gong, X., Bahnan, W., Barreras, H., Hernandez, R., Li, Z., Plano, G. V., & Schesser, K. (2019). Chromosomally-Encoded Yersinia pestis Type III Secretion Effector Proteins Promote Infection in Cells and in Mice. Frontiers in Cellular and Infection Microbiology, 9(Issue). doi:10.3389/fcimb.2019.00023
  More info

  Yersinia pestis, the causative agent of plague, possesses a number of virulence mechanisms that allows it to survive and proliferate during its interaction with the host. To discover additional infection-specific Y. pestis factors, a transposon site hybridization (TraSH)-based genome-wide screen was employed to identify genomic regions required for its survival during cellular infection. In addition to several well-characterized infection-specific genes, this screen identified three chromosomal genes (y3397, y3399, and y3400), located in an apparent operon, that promoted successful infection. Each of these genes is predicted to encode a leucine-rich repeat family protein with or without an associated ubiquitin E3 ligase domain. These genes were designated Yersinia leucine-rich repeat gene A (ylrA), B (ylrB), and C (ylrC). Engineered strains with deletions of y3397 (ylrC), y3399 (ylrB), or y3400 (ylrA), exhibited infection defects both in cultured cells and in the mouse. C-terminal FLAG-tagged YlrA, YlrB, and YlrC were secreted by Y. pestis in the absence but not the presence of extracellular calcium and deletions of the DNA sequences encoding the predicted N-terminal type III secretion signals of YlrA, YlrB, and YlrC prevented their secretion, indicating that these proteins are substrates of the type III secretion system (T3SS). Further strengthening the connection with the T3SS, YlrB was readily translocated into HeLa cells and expression of the YlrA and YlrC proteins in yeast inhibited yeast growth, indicating that these proteins may function as anti-host T3S effector proteins.
• Xu, B., Liu, C., Qian, L., Qu, Y., Su, W., Xu, J., & Zhao, J. (2019). Statistical modelling outcome of in vitro fertilization and intracytoplasmic sperm injection: A single centre study. Combinatorial Chemistry and High Throughput Screening, 22(Issue 4). doi:10.2174/1386207322666190404145448
  More info

  Background: Assisted reproductive techniques (ART) have been extensively used to treat infertility. Inaccurate prediction of a couple’s fertility often leads to lowered self-esteem for patients seeking ART treatment and causes fertility distress. Objective: This prospective study aimed to statistically analyze patient data from a single reproductive medical center over a period of 18 months, and to establish mathematical models that might facilitate accurate prediction of successful pregnancy when ART are used. Methods: In the present study, we analyzed clinical data prospectively collected from 760 infertile patients visiting the second Affiliated Hospital and Yuying Children’s Hospital of Wenzhou Medical University between June 1, 2016 and December 31, 2017. Various advanced statistical methods, including broken-line regression, were employed to analyze the data. Results: Age remained the most important factor affecting the outcome of IVF/ICSI. Using the broken-line regression model, the fastest clinical pregnancy declining age was between 25 and 32. Female infertility type was found to be a key predictor for the number of good-quality embryos and successful pregnancy, along with the antral follicle count (AFC), total number of embryos, recombinant follicle stimulating hormones (rFSH) dosage, estradiol (E2) on the trigger day, and total number of oocytes retrieved. rFSH dosage was also significantly associated with the number of oocytes retrieved and the number of frozen embryos. Conclusion: The fastest clinical pregnancy declining age is ranged between 25 and 32, and female infertility type is evidenced as another key predictive factor for the cumulative outcome of ART.
• Qian, L., & Wang, S. (2017). Subject-wise empirical likelihood inference in partial linear models for longitudinal data. Computational Statistics and Data Analysis, 111(Issue). doi:10.1016/j.csda.2017.02.001
  More info

  In analyzing longitudinal data, within-subject correlations are a major factor that affects statistical efficiency. Working with a partially linear model for longitudinal data, a subject-wise empirical likelihood based method that takes the within-subject correlations into consideration is proposed to estimate the model parameters. A nonparametric version of the Wilks Theorem for the limiting distribution of the empirical likelihood ratio, which relies on a kernel regression smoothing method to properly centered data, is derived. The estimation of the nonparametric baseline function is also considered. A simulation study and an application are reported to investigate the finite sample properties of the proposed method. The numerical results demonstrate the usefulness of the proposed method.
• Li, Y., & Qian, L. (2014). Likelihood ratio test for a piecewise continuous Weibull model with an unknown change point. Journal of Mathematical Analysis and Applications, 412(Issue 1). doi:10.1016/j.jmaa.2013.10.069
  More info

  In this paper, we consider the likelihood ratio test for the scale and shape parameters in a piecewise continuous Weibull model with an unknown change point. Under the null hypothesis of no change in scale and shape parameters, we derive that the likelihood ratio process converges weakly to the squared Euclidian norm of a weighted mean zero Gaussian vector process. © 2013 Elsevier Inc.
• Zhang, W., Qian, L., & Li, Y. (2014). Semiparametric sequential testing for multiple change points in piecewise constant hazard functions with long-term survivors. Communications in Statistics: Simulation and Computation, 43(Issue 7). doi:10.1080/03610918.2012.742106
  More info

  In this article, we consider parameter estimation in the hazard rate with multiple change points in the presence of long-term survivors. We combine two methods: maximum likelihood based and martingale based, to estimate the change points in the hazard rate for right censored survival data that accounts for long-term survivors. A simulation study is carried out to compare the performance of estimators. The method is applied to analyze two real datasets. © 2014 Copyright Taylor and Francis Group, LLC.
• Li, Y., Qian, L., & Zhang, W. (2013). Estimation in a change-point hazard regression model with long-term survivors. Statistics and Probability Letters, 83(Issue 7). doi:10.1016/j.spl.2013.03.026
  More info

  This paper estimates the change-point for a piecewise hazard regression model in the presence of right censoring and long-term survivors. The maximum likelihood estimators of the change point and other parameters are shown to be consistent. The proposed method is illustrated through analyzing the kidney infection recurrence data. © 2013 Elsevier B.V.
• Long, H., & Qian, L. (2013). Nadaraya-Watson estimator for stochastic processes driven by stable Lévy motions. Electronic Journal of Statistics, 7(Issue 1). doi:10.1214/13-ejs811
  More info

  We discuss the nonparametric Nadaraya-Watson (N-W) estimator of the drift function for ergodic stochastic processes driven by α-stable noises and observed at discrete instants. Under geometrical mixing condition, we derive consistency and rate of convergence of the N-W estimator of the drift function. Furthermore, we obtain a central limit theorem for stable stochastic integrals. The central limit theorem has its own interest and is the crucial tool for the proofs. A simulation study illustrates the finite sample properties of the N-W estimator.
• Bartra, S. S., Gong, X., Lorica, C. D., Jain, C., Nair, M. K., Schifferli, D., Qian, L., Li, Z., Plano, G. V., & Schesser, K. (2012). The outer membrane protein A (OmpA) of Yersinia pestis promotes intracellular survival and virulence in mice. Microbial Pathogenesis, 52(Issue 1). doi:10.1016/j.micpath.2011.09.009
  More info

  The plague bacterium Yersinia pestis has a number of well-described strategies to protect itself from both host cells and soluble factors. In an effort to identify additional anti-host factors, we employed a transposon site hybridization (TraSH)-based approach to screen 10 5 Y. pestis mutants in an in vitro infection system. In addition to loci encoding various components of the well-characterized type III secretion system (T3SS), our screen unambiguously identified ompA as a pro-survival gene. We go on to show that an engineered Y. pestis ΔompA strain, as well as a ΔompA strain of the closely related pathogen Yersinia pseudotuberculosis, have fully functioning T3SSs but are specifically defective in surviving within macrophages. Additionally, the Y. pestis ΔompA strain was out competed by the wild-type strain in a mouse co-infection assay. Unlike in other bacterial pathogens in which OmpA can promote adherence, invasion, or serum resistance, the OmpA of Y. pestis is restricted to enhancing intracellular survival. Our data show that OmpA of the pathogenic Yersinia is a virulence factor on par with the T3SS. © 2011 Elsevier Ltd.
• Qian, L. (2012). The Fisher information matrix for a three-parameter exponentiated Weibull distribution under type II censoring. Statistical Methodology, 9(Issue 3). doi:10.1016/j.stamet.2011.08.007
  More info

  This paper considers the three-parameter exponentiated Weibull family under type II censoring. It first graphically illustrates the shape property of the hazard function. Then, it proposes a simple algorithm for computing the maximum likelihood estimator and derives the Fisher information matrix. The latter is represented through a single integral in terms of the hazard function; hence it solves the problem of computational difficulty in constructing inferences for the maximum likelihood estimator. Real data analysis is conducted to illustrate the effect of the censoring rate on the maximum likelihood estimation. © 2011 Elsevier B.V.
• Liu, Z., & Qian, L. (2010). Changepoint estimation in a segmented linear regression via empirical likelihood. Communications in Statistics: Simulation and Computation, 39(Issue 1). doi:10.1080/03610910903312193
  More info

  For a segmented regression system with an unknown changepoint over two domains of a predictor, a new empirical likelihood ratio statistic is proposed to test the null hypothesis of no change. Under the null hypothesis of no change, the proposed test statistic is shown empirically to be Gumbel distributed with robust location and scale estimators against various parameter settings and error distributions. A power analysis is conducted to illustrate the performance of the test. Under the alternative hypothesis with a changepoint, the test statistic is utilized to estimate the changepoint between the two domains. A comparison of the frequency distributions between the proposed estimator and two parametric methods indicates that the proposed method is effective in capturing the true changepoint.
• Wang, S., Qian, L., & Carroll, R. J. (2010). Generalized empirical likelihood methods for analyzing longitudinal data. Biometrika, 97(Issue 1). doi:10.1093/biomet/asp073
  More info

  Efficient estimation of parameters is a major objective in analyzing longitudinal data. We propose two generalized empirical likelihood-based methods that take into consideration within-subject correlations. A nonparametric version of the Wilks theorem for the limiting distributions of the empirical likelihood ratios is derived. It is shown that one of the proposed methods is locally efficient among a class of within-subject variance-covariance matrices. A simulation study is conducted to investigate the finite sample properties of the proposed methods and compares them with the block empirical likelihood method by You et al. (2006) and the normal approximation with a correctly estimated variance-covariance. The results suggest that the proposed methods are generally more efficient than existing methods that ignore the correlation structure, and are better in coverage compared to the normal approximation with correctly specified within-subject correlation. An application illustrating our methods and supporting the simulation study results is presented. © 2010 Biometrika Trust.
• Diaz, N., Lizardi, H., Qian, L., & Liu, Z. (2008). The relationship among child maltreatment, parental bonding, and a lifetime history of major depressive disorder in latino college students. Journal of Aggression, Maltreatment and Trauma, 17(Issue 2). doi:10.1080/10926770802344885
  More info

  This study examined the relationship among child maltreatment, parental bonding, and a lifetime history of major depressive disorder (MDD) in a sample of 119 Latino students. Forty-five students reported a lifetime history of MDD and 74 reported not having a lifetime history of MDD. The results indicated that emotional abuse and maternal overprotection were significantly associated with having a lifetime history of MDD. The findings support the importance of examining these factors among depressed Latinos. Future research should continue to explore both the role of child maltreatment and parental bonding in relation to a lifetime history of MDD in this population. © 2008 by The Haworth Press. All rights reserved.
• Heilmayer, O., Digialleonardo, J., Qian, L., & Roesijadi, G. (2008). Stress tolerance of a subtropical Crassostrea virginica population to the combined effects of temperature and salinity. Estuarine, Coastal and Shelf Science, 79(Issue 1). doi:10.1016/j.ecss.2008.03.022
  More info

  The combination of salinity and temperature has synergistic effects on virtually all aspects of the biology of estuarine organisms. Of interest were site-specific characteristics in the response of the eastern oyster, Crassostrea virginica, from the St. Lucie River Estuary to the interactive effects of temperature and salinity. This estuary, one of the largest on the central east coast of Florida, is strongly influenced by anthropogenic modifications due to management needs to control the patterns of freshwater flow in the St. Lucie River watershed. Crassostrea virginica is designated a valued ecosystem component for monitoring the health of this estuary. Our approach used a multidimensional response surface design to study the effects of temperature and salinity on sublethal measures of oyster performance: (1) body condition index as an overall indicator of bioenergetic status and (2) the RNA/DNA ratio as a biochemical indicator of cellular stress. The results showed that there was a greater ability to withstand extreme salinity conditions at lower temperatures. However, there were no site-specific attributes that differentiated the response of the St. Lucie Estuary population from populations along the distribution range. Condition index was a less variable response than the RNA/DNA ratio, and the final models for mean condition index and the RNA/DNA ratios explained 77.3 and 35.8% of the respective variances. © 2008 Elsevier Ltd. All rights reserved.
• Qian, L., & Ryu, S. (2006). Estimating tree resin dose effect on termites. Environmetrics, 17(Issue 2). doi:10.1002/env.761
  More info

  The aim of this study is to determine the effectiveness of a natural tropical tree resin in controlling termites thus providing protection from their destruction. Tree resin from the bark of tropical trees offers the potential for this protection. Termites were fed by filter paper soaked in tropical tree resin dissolved in a solvent at different concentrations for 15 days. The number of termites still alive on each day was observed and recorded. In this paper, we use four types of statistical models: Partially linear model, piecewise linear model, cubic smooth spline and mixed effect model to analyze the termite data. The results show that tropical tree resin, particularly at a higher concentration of 10mg is significantly more effective in killing termites. We show that two dishes under 10 mg of tree resin were mistaken since the data for these two dishes are shown insignificantly different from data under 5mg. The partially linear model shows that there is non-linear (piecewise linear) time effect. Both piecewise linear model and cubic spline smoothing show that the most effective period is the first week. The non-parametric smoothing, cubic spline, and piecewise linear model are not significantly different. Mixed effect model is consistent with partial linear model and piecewise linear model. The estimated treatment effect is time varying with a change point at day 7. Therefore, we suggest the piecewise linear model as the final simplest one for prediction. This model fits the data with adjusted R2 = 93.7 per cent and shows that on average, 10 mg is 68.9 per cent more efficient than 5 mg in killing termites during the first week. Copyright © 2005 John Wiley & Sons, Ltd.
• Koul, H. L., Qian, L., & Surgailis, D. (2003). Asymptotics of M-estimators in two-phase linear regression models. Stochastic Processes and their Applications, 103(Issue 1). doi:10.1016/s0304-4149(02)00185-0
  More info

  This paper discusses the consistency and limiting distributions of a class of M-estimators in two-phase random design linear regression models where the regression function is discontinuous at the change-point with a fixed jump size. The consistency rate of an M-estimator r̂n for the change-point parameter r is shown to be n while it is n1/2 for the coefficient parameter estimators, where n denotes the sample size. The normalized M-process is shown to be uniformly locally asymptotically equivalent to the sum of a quadratic form in the coefficient parameter vector and a jump point process in the change-point parameter, in probability. These results are then used to obtain the joint weak convergence ofthe M-estimators. In particular, n(r̂n - r) is shown to converge weakly to a random variable which minimizes a compound Poisson process, a suitably standardized coefficient parameter M-estimator vector is shown to be asymptotically normal, and independent of n(r̂n - r). © 2002 Elsevier Science B.V. All rights reserved.
• Qian, L., & Correa, J. A. (2003). Estimation of Weibull parameters for grouped data with competing risks. Journal of Statistical Computation and Simulation, 73(Issue 4). doi:10.1080/0094965021000033431
  More info

  We consider the problem of estimating Weibull parameters for grouped data when competing risks are present. We propose two simple methods of estimation and derive their asymptotic properties. A Monte Carlo study was carried out to evaluate the performance of these two methods.
• Koul, H. L., & Qian, L. (2002). Asymptotics of maximum likelihood estimator in a two-phase linear regression model. Journal of Statistical Planning and Inference, 108(Issue 1-2). doi:10.1016/s0378-3758(02)00273-2
  More info

  This paper considers two-phase random design linear regression models with arbitrary error densities and where the regression function has a fixed jump at the true change-point. It obtains the consistency and the limiting distributions of maximum likelihood estimators of the underlying parameters in these models. The left end point of the maximizing interval with respect to the change point, herein called the maximum likelihood estimator r̂n of the change-point parameter r, is shown to be n-consistent and the underlying likelihood process, as a process in the standardized change-point parameter, is shown to converge weakly to a compound Poisson process. This process obtains maximum over a bounded interval and n(r̂n - r) converges weakly to the left end point of this interval. These results are different from those available in the literature for the case of the two-phase linear regression models when jump sizes tend to zero as n tends to infinity. © 2002 Elsevier Science B.V. All rights reserved.
• Austin, D. F., Kitajima, K., Yoneda, Y., & Qian, L. (2001). A putative tropical American plant, Ipomoea nil (Convolvulaceae), in pre-Columbian Japanese art. Economic Botany, 55(Issue 4). doi:10.1007/bf02871714
  More info

  The Heike Nôkyô, Japanese scrolls of Buddhist sutras created in 1164 AD, includes illustrations of an Ipomoea that has long been identified by Japanese scholars as I. nil. What makes this occurrence of I. nil in pre-Columbian Japan remarkable is that all of its closest relatives are American plants. We give a synopsis of the history of this economically important species. Then, using cladistic analysis, we show the relationships of I. nil to I. eriocalyx, I. hederacea, I. indica, I. laeta, I. lindheimeri, I. meyeri, and I. pubescens. Six of these eight species in Ipomoea series Heterophyllae are endemic to the New World. Ipomoea indica is pantropical, and may be carried by ocean currents. We offer four hypotheses as to how this putatively tropical American species may have arrived in Asia: 1) Ipomoea nil was introduced through long-distance dispersal by animals; 2) Ipomoea nil was introduced by humans in a pre-Columbian context; 3) The Ipomoea in the Heike Nôkyô scrolls does not represent I. nil, but a different native Asian species; and 4) Ipomoea nil was introduced during post-Columbian times by Europeans. There are problems with accepting any of these possible alternatives.
• Qian, L., & Wang, S. (2001). Bias-corrected heteroscedasticity robust covariance matrix (sandwich) estimators. Journal of Statistical Computation and Simulation, 70(Issue 2). doi:10.1080/00949650108812114
  More info

  Two simple bias-corrected sandwich estimators are proposed for the covariance of the least squared coefficient estimator in the linear models. These estimators are unbiased with homoscedastic errors and are shown to be robust against moderate deviations from the homoscedasticity assumption. Simulation results suggest that one of the proposed estimators produces at most a small bias but with an increased variance while the other produces a smaller mean squared error than the classical estimators such as those of Hinkley (1977), White (1980), and Furno (1997) in both cases of homoscedastic and heteroscedastic errors.
• Qian, L. (1998). On maximum likelihood estimators for a threshold autoregression. Journal of Statistical Planning and Inference, 75(Issue 1). doi:10.1016/s0378-3758(98)00113-x
  More info

  For a stationary ergodic self-exciting threshold autoregressive model with single threshold parameter, obtained the consistency and limiting distribution of the least-squares estimator for the underlying true parameters. In this paper, we derive the similar results for the maximum likelihood estimators of the same model under some regularity conditions on the error density, not necessarily Gaussian.
• Qian, L. (1996). Minimum distance estimators for random coefficient autoregressive models. Statistics and Probability Letters, 29(Issue 3). doi:10.1016/0167-7152(95)00180-8
  More info

  This paper discusses a class of minimum distance (MD) estimators for a class of pth-order random coefficient autoregressive (RCAR(p)) models. These estimators are defined via certain weighted empiricals as in Koul (1986). The class of estimators considered includes the least absolute deviation estimator and an analogue of the Hodges-Lehmann estimator. The paper contains a proof of the asymptotic normality of these estimators and a simulation study. It is observed that the RCAR(2) model with the Hodges-Lehmann type estimator fits the Canadian lynx data at least as well as with the least square estimator.