Mathematics Department, University of Maryland, College Park, MD 1967-1999
ADVISOR: K. DANIEL
Brooks, Lowell (6/70) Linear Programs for the Solutions of Discrete Dynamic Programming and Markov Renewal Programming Problems
ADVISOR: M. FREIDLIN
Sowers, Richard (5/91) New Asymptotic Results for Stochastic Partial Differential Equations
Carmona, Sara (5/92) A Large Deviations Principle and Wave Front Propagation for a Reaction-Diffusion Equation
Dunyak, James (5/94) Probabilistic Methods for Diffusions in Regions with Many Small Holes
Vilarrubi, Roberto (12/94) Large Deviations Results for Some Stochastic Partial Differential Equations
Rowe, Errol (5/96) Probabilistic Approach to a class of PDE systems
Pfeiffer, Ruth (8/98) Statistical Problems for Stochastic Processes with Hysteresis
ADVISOR: A. KAGAN
Tzavelas, George (8/94) Parameter Estimation: Quasi-Likelihood, Generalized Linear Models, Semiparametric Models and Exponential Families
ADVISOR: B. KEDEM
Reed, George (12/83) Some Properties and Applications of Higher Order Crossings
Martin, Donald E. (12/90) Estimation of the Minimal Period of Periodically Correlated Sequences
Lopes, Silvia (12/91) Spectral Analysis in Frequency Modulated Models
Pavlopoulos, Haralabos (12/91) Statistical Inference for Optimal Thresholds
Troendle, James (12/91) An Iterative Filtering Method of Frequency Detection in a Mixed Spectrum Model
Li, Ta-Hsin (8/92) Multiple Frequency Estimation in Mixed-Spectrum Time Series by Parametric Filtering
Fokianos, Kostantinos (8/96) Categorical Time Series: Prediction and Control
Barnett, John T. (8/96) Zero-Crossings of Some Non-Gaussian Processes with Application to Estimation
De Oliveira, Victor (8/97) Prediction in Some Classes of Non-Gaussian Random Fields
Jeffries, Neal (5/98) Logistic Mixtures of Generalized Linear Models in Time Series
Kozintsev, Boris (8/99) Computations With Gaussian Random Fields
ADVISOR: T. LEE
Torcaso, Fred (8/98) Wave Propagation for Systems of KPP Equations in Random Media
ADVISOR: P. MIKULSKI
Kaplan, Harold M. (6/67) Large-Sample Tests of Expectations in the Presence of Nuisance Parameters
Kalish, George (6/69) The Asymptotic Efficiency of the Smirnov Test
Yu, Chin Shih (8/69) On Upper Bounds of Asymptotic Power and on Pitman Efficiencies of Kolmogorov and Smirnov Tests
Archambault, William (12/72) On Bounds for the Asymptotic Power and on Pitman Efficiencies for the Cramer-von-Mises Tests
Wieand, Harry (Sam) (12/74) On a Condition Under Which the Pitman and Bahadur Approaches to Efficiency Coincide
Weiner, Kenneth (5/75) Modified Approximate Efficiencies for General Location Problems
Monsour, Michael (8/86) Optimality and Other Asymptotic Properties of the Maximum Likelihood Estimator in the First Order Autoregressive Process
ADVISOR: E. SLUD
Chambers, Daniel (8/83) Central and Functional Central Limit Theorems for Functionals of Gaussian Processes
Koutsoukos, Antonis (8/90) Probabilities of Moderate and Large Deviations of Test Statistics and Estimators in the Presence of Nuisance Parameters
Vonta, Filia (8/92) Efficient Estimation of a Structural Parameter in a Non-Proportional Hazards Model in the Two-Sample Problem
Kopylev, Leonid (8/97) On Estimation of the Marginal Survival Function in the Cox Model
ADVISOR: P. SMITH
Lakatos, Edward (12/78) Undiminished Residual Effects Designs and Their Applications to Survey Sampling
Hurtado-Donaldson, Ana (5/88) Nonparametric Estimation in a Survival/Sacrifice Experiment
Myers, Margaret (8/88) Robustness of Design in Misspecified Logistic Re- gression
Chen, Sy-Mien (12/90) Robust Tests in Statistical Quality Control
Graubard, Barry I. (8/91) Statistical Methods for the Analysis of Complex Survey Data with Biomedical Applications
Chung, Carol (12/97) A Central Limit Theorem for Spatial Regression Based on Generalized Estimating Equations Applications
ADVISOR: R. SYSKI
Shachtman, Richard H. (8/68) Stieltjes Stochastic Integrals
Blake, Louis (8/69) The Preservation of Regularity of Conditional Proba- bilities
Kolmus, Peter (8/75) Extension Theorems for Choquet Capacities with Applications in Probability Theory
Bushar, Harry (8/76) Continuous Parameter Markov Chains, Classification of Boundary Points and Use of Local Time
Harrington, David (8/76) Limit Theorems for Continuous Time Markov Branching Processes with Varying and Random Environments
Rosen, Julie (12/81) The Fortet Integral with Respect to a Martingale
Cai, Haiyan (5/88) On Reviving Markov Processes and Applications
Liu, Ning (5/95) Decomposition Theorems for Standard Processes
ADVISOR: C. Z. WEI
Chan, Ngai Hang (8/85) Asymptotic Inference of Nonstationary Autore-gressive Processes
Winnicki, Jan (5/86) A Unified Estimation Theory for the Branching Process with Immigration
Guo, Meihui (8/89) Inference for Nonlinear Time Series
Lee, Sangyeol (8/91) Testing Gaussianality of Time Series
Karagrigoriou, Alexandros (8/92) Asymptotic Efficiency of Model Selection Procedures in Time Series
ADVISOR: G. YANG
Fleming, Thomas (12/76) Non-Parametric Estimation for Non-Time-Homogeneous Markov Processes in the Problem of Competing Risks
Rubinstein, Lawrence (8/78) Nonparametric Estimation of the Survival Function for Censored Data
Conner, Teresa (12/81) A Heteroscedastic Model Arising from Dependence Between the Mean and the Variance
Chang, Myron (5/84) Nonparametric Estimation for Doubly Censored Data
Foster, Dean (8/88) Conditional Least Squares for Semi-Martingales
Lee, Chin San (8/89) Parameter Estimation in Branching Processes with Application to Tumor Growth
Wang, Wenyu (5/90) Statistical Inference for Aggregated Markov Processes
Chen, Gang (8/92) A Conditional Bootstrap Procedure and its Asymptotic Accuracy: A Nonparametric Approach
Xu, Jian-Lun (8/93) Nonparametric Estimation of a Distribution Function in Biased Sampling Models
Admission to candidacy for the doctoral degree is granted by the Graduate School upon the recommendation of the MATH Graduate Committee. A student must be admitted to candidacy within five years after admission to the doctoral program and at least six months before the date on which the doctoral degree will be conferred. Before a student applies for admission to candidacy he or she must have:
It is the responsibility of the student to submit an application for admission to candidacy to the Graduate Director when all the requirements for candidacy have been fulfilled. Application forms may be obtained at the MATH office. All work at other institutions offered in partial fulfillment of the requirements for the doctoral degree must be submitted with the application for admission to candidacy. Official transcripts of the work must be on file in the Graduate School. The student must complete his or her program for the degree, including the foreign language examination, dissertation, and final examination (defense), during the four year period after admission to candidacy.
The Oral Candidacy Examination: The candidacy examination is an oral examination which serves as a test of the detailed preparation of a student in the area of specialization, and seeks to discover if he or she has a deep enough understanding to read the relevant research literature in the field and the skills to carry out the research for the dissertation. The examination is usually taken before a student embarks on serious dissertation research. The examination assumes further advanced course work beyond that required for the qualifying exams. (Sample programs of such advanced course work in various fields may be found here.) It shall follow the guidelines listed below.
Planning the Exam: To plan the examination, the student, with the help and approval of the prospective dissertation advisor, must prepare a prospectus for the examination. This prospectus defines the primary and related areas to be covered in the examination. These areas should be identified by course citations, literature citations, tables of contents, or other appropriate means. The prospectus should be filed with the Graduate Office before the examination is scheduled, and should also record the proposed format for the examination. Typical formats for the examination are either a seminar-type presentation by the student (or possibly two such talks) on one or more recent research papers, followed by questions from the committee on the presentation and related background material, or else a more traditional oral examination on subjects or courses listed in the prospectus.
Examination Committee: The examination committee is appointed by the Graduate Director (or if the Graduate Director is unavailable for an extended period, his or her designated representative) upon recommendation of the student's prospective dissertation advisor. The Graduate Director may if necessary consult with one or more field committee chairs in the area of specialization. The examination committee must consist of at least three members, at least one (usually the prospective dissertation advisor) representing the area in which the student plans to specialize. Usually all three of these will be faculty members from the Mathematics Department, but when there is a good academic reason, the student can petition the Graduate Committee to allow one to be from a related department (such as physics or computer science) or an outside institution (such as another university, NASA, NIH, NIST, NCHS, etc.). Disputes regarding the makeup of the examination committee will be referred to the Graduate Committee. Each committee member must agree to abide by the prospectus for the examination.
Possible Outcomes: Upon completion of the examination, the examination committee decides to pass, fail, or defer a decision on the student. In the last-named case, the manner in which the decision is to be resolved must be specified in the report of the committee. The distinction between "fail" and "defer a decision" is based on the committee's evaluation of the probability of successful completion of the Ph.D. degree.
Repeating the Exam: Upon failure, the Candidacy Examination may be repeated only once. Exceptions to this rule are granted only under extraordinary circumstances and upon petition to the Graduate Committee.
The Mathematical Statistics Program offers M.A. and Ph.D. degrees in statistics and probability theory with areas of faculty specialization including stochastic processes, statistical decision theory, biostatistics, stochastic modeling, nonparametric inference, multivariate analysis, categorical data, time series analysis and large sample theory. Students may pursue a program of study emphasizing either theory or applications by appropriate choice of coursework and research topics. The program has been designed with sufficient flexibility to accommodate the student's background and interests.
Academic matters relating to the Mathematical Statistics Program are determined by the Statistics faculty of the Mathematics Department. The Department administers graduate programs in Mathematics (MATH) and Mathematical Statistics (STAT), and also cooperates with the program in Applied Mathematics and Scientific Computation (AMSC). Administrative support for all three programs is provided by the Department's Office of Graduate Studies. In particular, all teaching assistantships and most fellowships for students in the three programs are handled by the Mathematics Graduate Office.
During the first year a graduate student has the privilege of transferring among the three related graduate programs of MATH, STAT, and AMSC. After the first year, switching between programs is possible but not automatic, and requires approval of the programs involved.
Some general regulations of the Graduate School are listed in this brochure as well as specific policies of the Department. These policies should be carefully considered by all graduate students in planning their work towards an advanced degree. Additional information is available in the Office of Graduate Studies and in such publications as the Graduate Catalog and the Schedule of Classes.
Every student is expected to meet with an advisor each semester. Upon admission new students should follow the advising directions of the Departmental letter of admission (except those who will be graduate teaching assistants). The new teaching assistants complete advising and registration during the one week mandatory orientation program that takes place in August, the week before the start of classes. For currently enrolled students, registration takes place either electronically or through the Office of Graduate Studies every fall and spring for the following semester.
All Statistics Program faculty members can advise graduate students on their program and selection of courses. However, for the purpose of coordination and course planning it is expected that graduate students in Statistics will consult with the Statistics Program Director about their plans for immediate course selection and expected registration for following semesters. New graduate students will usually be advised by the Statistics Program Director.
The advisor and the student work together to formulate the appropriate course of study. The program should combine core material in statistics and probability, supporting material in mathematics and/or areas of application of statistics, and more specialized study in areas of particular interest to the student. There are no specific course requirements. However, a narrow, over-specialized program is undesirable, since statisticians must be able to apply their knowledge to a variety of problems and must have a wide range of skills at their disposal. The program is subject to the approval of the Director of the Statistics Program.
Core Courses: All students should plan to take STAT 650 and STAT 700-701. In addition, those with a weak background in probability and statistics should take STAT 410 in their first semester at Maryland.
M.A.--Thesis Option: In addition to the core courses, students elect other courses in statistics, mathematics or areas of application. This enables the student to set up an individualized program of study in applied statistics, mathematical statistics or applied probability. The student should plan on beginning thesis research in the second year.
M.A.--Non Thesis Option: These students take the departmental written comprehensive examination in statistics, probability, and a third area of statistics or mathematics. The program of study should include the core courses, a course sequence for the third part of the written examination, and other courses in statistics, mathematics or applied areas to complete the program. In addition, candidates choosing the non-thesis option for the M.A. must prepare a scholarly paper.
Ph.D.-- The doctoral student's program usually includes STAT 600-601-650, STAT 700-701 and a mathematics sequence. This is the core material for the three part Ph.D. written examination. In addition, doctoral students should plan on taking some more advanced courses, usually at least some subset of STAT 740-741 (Linear Statistical Models), STAT 710 (Advanced Statistics), and STAT 750 (Multivariate Analysis). Advanced students often take independent reading courses in their areas of research in addition to or instead of formal course work. Participation in the probability and statistics seminar and statistics workshop is required of all who plan to write a Ph.D. dissertation. In addition, these students must give a presentation in some area of current research in the field.
Any course may be repeated and the grade in the repeated course replaces the original grade in determining the overall average. As long as the overall average is at least B at the time of receiving the degree, grades of D, F and I may stand, but D and F count as 0 quality points in computing averages, and courses in which these grades are received cannot be used to fulfill degree requirements.
The Schedule of Classes should be consulted for pertinent dates for adding and dropping classes.
A full time graduate student must carry a combination of courses that adds up to at least 48 units each semester (excluding the summer sessions). For graduate assistants this requirement is reduced to a minimum of 24 units. A unit is defined as follows:
All 400 level courses: 4 units per credit hour.
All 600/700 level courses: 6 units per credit hour (except 799) 799: 12 units per credit hour 899: 18 units per credit hour
Each professor has an individual section number for a reading or research course. This section number is available in the Office of Graduate Studies. Students registering for an RIT (STAT 689), a reading course (STAT 698 or 798) or any 799/899 course should obtain the correct section number from that office.
Students are expected to make steady progress toward their degrees. For the M.A. degree, all requirements must be completed within five years from the date of admission. A student admitted to a Ph.D. program must be admitted to candidacy within five years from the date of admission. After admission to candidacy all requirements for the Ph.D. degree must be completed within four additional years. Minimal continuous registration is required of all students who have been admitted to candidacy for the Ph.D. degree.
Graduate teaching assistantships constitute the main form of financial aid offered by the Department. In addition to a stipend, graduate assistants receive a tuition scholarship for up to ten credits per semester and are eligible for health care benefits. See here for further information on duties of graduate assistants and renewal policies for fellowships and assistantships.
A student who receives a Master's degree in Mathematical Statistics should demonstrate a general understanding of the main branches of the subject and must have shown a high level of scholarship and ability. Two options are available for this degree: the M.A. with thesis and the M.A. without thesis.
Residence Requirements: A full-time student must have two semesters in residence, a part-time student four semesters. All requirements for the M.A. degree must be completed within a period of five years.
Transfer of Credit: Up to 6 credits of graduate level work taken at another regionally accredited institution is permitted under the following provisions:
Diploma Application: Applications for Diploma should be made at the Records Office, Room 1101, Mitchell Building, early in the semester in which the degree is expected. The deadline for application is listed in the Schedule of Classes.
Approved Program Form: A student who has applied for a diploma must complete the Approved Program form obtained from the Office of Graduate Studies before the deadline listed in the Schedule of Classes. This form is returned to the Administrator of the Graduate Program who will forward it to the Graduate School.
Grade Point Average: The student must maintain an average of B or better in all courses taken, not just those listed in the Approved Program. For this purpose, the grades of D, F, and I count as 0 quality points, and courses with these grades cannot be used for degree requirements.
Incomplete: Any grade of incomplete in a course listed in the Approved Program must be removed.
In addition to satisfying the requirements applicable to all M.A. candidates, the student must have:
Thesis: The M.A. thesis should represent a meaningful piece of independent work which has some novel features, for example, the detailed working out of the application of a general theory or method to some particular case or cases of interest. It must be prepared in the form required by the Graduate School. Each member of the final oral committee must receive a legible typed copy at least one week before the final oral examination. Two copies of the thesis must be delivered to the Graduate School after the final oral examination and before the deadline specified in the Schedule of Classes.
Nomination of Thesis or Dissertation Committee Form: This form, obtained from the Office of Graduate Studies, must be completed two months prior to the date of the final oral and in keeping with the deadline listed in the Schedule of Classes. It should be completed in conjunction with the student's thesis advisor and returned to the Administrator of the Graduate Program who will forward it to the Graduate School. This will generate the Report of Examining Committee form sent from the Graduate School to the Statistics Director to be taken to the final oral examination. It should be signed by all members of the thesis committee and returned to the Graduate School. There is also an equivalent internal form. The student will be examined on the thesis and related topics at the discretion of the examiners. All pertinent information concerning this oral examination should be given to the Office of Graduate Studies two weeks prior to the examination. The information will then be posted, as the examination is open to the public.
In addition to satisfying the requirements applicable to all M.A. candidates, the student must have:
Scholarly paper: The student must complete an acceptable scholarly paper of an expository nature. Normally, the topic shall be related to an advanced course or seminar taken in partial fulfillment of the course requirements for the degree. The topic shall normally be agreed upon with the professor in the course, who shall become the student's advisor. If the paper is not written in connection with a course, some other appropriate faculty member may approve the topic and become the advisor. A second reader shall be appointed by the Statistics Program Director and both readers must approve in order for the paper to be accepted. A neat copy of the final approved version shall be provided for the Office of Graduate Studies files. The scholarly paper shall be based on substantial use of at least two sources, including one journal article. The paper must include an abstract and references to all literature used.
Final Oral Examination: The final oral examination shall consist of a presentation of the material in the scholarly paper, plus questioning by the examiners based on the paper and whatever material in the approved M.A. program that has not been covered by the written examination. The examining committee shall consist of the two readers of the scholarly paper.
To receive the Ph.D. degree in mathematical statistics a student must display a high level of scholarship shown by the ability to do original research and should possess a broad knowledge of major fields of the subject. It is not necessary to obtain a master's degree before obtaining the doctorate.
Residence Requirements: The equivalent of at least three full years of graduate study is required, of which at least one must be in residence at the University of Maryland campus. At least 18 hours of course work must be taken at the University of Maryland, plus 12 hours of research at the Ph.D. level.
Minimum Requirements: In order to receive a Ph.D. degree, the student must have:
Doctoral Candidate's Presentation: As a condition for Ph.D. candidacy, the student must make an oral presentation in an area of current research. The level of the presentation should demonstrate depth of knowledge, familiarity with research literature, and ability to write a doctoral dissertation on a topic related to the subject of the presentation. The subject matter will be determined by the student with the help of his prospective thesis advisor. An examining committee of three statistics faculty members is appointed by the Program Director. At the conclusion of the presentation, the committee judges the presentation as acceptable or unacceptable. The committee may question the student on other material, if they deem such questioning necessary to reach a judgement.
Approved Program: The entire course of study must constitute a unified program, approved by an advisor in the field of the student's major interest and by the Program Director.
Admission To Candidacy: Before petitioning for admission to candidacy, a student must have:
After fulfilling these requirements, the student should complete the Admission to Candidacy form available in the Office of Graduate Studies. This will be forwarded to the Graduate School.Dissertation: The dissertation must represent an original contribution to existing knowledge in mathematical statistics or probability. It must follow the form given in the manual which is available at the Copy Center at Reckord Armory. Two copies of the dissertation, three copies of the abstract, and two title pages must be submitted to the Graduate School on the proper paper as stated in the manual. The copies must be submitted to the Graduate School Records Office, after the Final Oral Examination but before the deadline listed in the Graduate School's Calendar of Important Dates. It is expected that the dissertation or some modification thereof will be submitted to a statistical, mathematical or scientific journal for publication.
Final Oral Examination: The final oral examining committee must consist of five members, one of whom is a regular member of the graduate faculty of a department other than mathematics. Each member of the committee must be given a copy of the dissertation at least two weeks prior to the examination.
The Nomination of Thesis or Dissertation Committee form is obtained from the Office of Graduate Studies and must be completed and returned to that office three months prior to the final oral and in accordance with the deadline listed in the Schedule of Classes. Details governing the structure of the committee are on the back of this form. This will generate the Report of Examining Committee form sent from the Graduate School to the Statistics Director which must be taken to the final oral, signed by all members of the committee and returned to the Graduate School. There is also an equivalent internal form.
All pertinent information concerning the oral examination should be given to the Office of Graduate Studies two weeks prior to the examination. The information will then be posted as this examination is open to the public.
The final oral examining committee will examine the candidate on the research work incorporated in the dissertation, review attainments and then vote on the candidate's qualifications for the degree. In order to justify a finding of failure, at least two negative votes must be cast.
RITs ("Research Interaction Teams") are informal groups designed to foster interaction between faculty, students, and postdocs, and to get students interested in current research. Most of them meet as informal seminars with active student participation (and in many cases, student organization as well).
Professor: Prof. Grace Yang
Prerequisites: Minimum background in Stat 410 and Stat 700 or equivalent (with consent of the instructor).
Survival analysis concerns the statistical theory and methods for the analysis of time-to-event data or lifetime data. Lifetime data are commonly studied by researchers from diverse scienti c elds including physics, biology, medicine, public health, actuaries, epidemiology, economics and engineering reliability. Measure of lifetime requires the knowledge of both its beginning and its end points. These end points are not always possible to measure. For example, typically a detector can record the time at which a neutron disintegrates, but not the time it is generated, and this results in a left-censored neutron lifetime. Just the opposite in clinical trials, the survival time of a patient after a treatment might be right-censored if the patient withdraws from the trial before his/her death or if the trial terminates before the patient dies. There are numerous other forms of incomplete lifetimes such as doubly censored, interval censored, randomly truncated and current status data. Sampling methods, sampling subjects, experimental designs and limitations in recording instruments are among the contributing factors to the incomplete data. Proper treatment of incomplete data is necessary for eliminating bias in data analysis. The course is mainly about the statistical analysis of incomplete lifetime data. The topics to be presented include stochastic modeling of censored and random-truncation data, parametric and nonparametric methods, the Kaplan-Meier estimator of a survival function, the Lynden-Bell estimator, construction of con dence bands, the Cox regression model, logistic models, the Fix-Neyman competing risks model, asymptotic statistical inference. Emphasis will be on statisical methods with examples drawn from applications.
Please note that solutions in the google drive below have been written to assist graders and may not be exhaustive. If you find any major errors, please contact the Math Graduate Office at . You must sign into your umd gmail to access the solutions.
https://drive.google.com/drive/folders/18t2o4bn7SAfqBMbDIJk-016jxaDinE-v?usp=sharing