No description available.

Die aktuelle Ausgabe des Newsletter für Freunde, Absolventen und Ehemalige der Technischen Universität Chemnitz.

Gene prioritization involves ranking genes by possible relevance to a disease of interest. This is important in order to narrow down the set of genes to be investigated biologically, and over the years, several computational approaches have been proposed for automat-ically prioritizing genes using some form of gene-related data, mostly using statistical or machine learning methods. Recently, Agarwal and Sengupta (2009) proposed the use of learning-to-rank methods, which have been used extensively in information retrieval and related fields, to learn a ranking of genes from a given data source, and used this approach to successfully identify novel genes related to leukemia and colon cancer using only gene expression data. In this work, we explore the possibility of combining such learning-to-rank methods with rank aggregation techniques to learn a ranking of genes from multiple heterogeneous data sources, such as gene expression data, gene ontology data, protein-protein interaction data, etc. Rank aggregation methods have their origins in voting theory, and have been used successfully in meta-search applications to aggregate webpage rankings from different search engines. Here we use graph-based learning-to-rank methods to learn a ranking of genes from each individual data source represented as a graph, and then apply rank aggregation methods to aggregate these rankings into a single ranking over the genes. The thesis describes our approach, reports experiments with various data sets, and presents our findings and initial conclusions.

Die aktuelle Ausgabe des Newsletter für Freunde, Absolventen und Ehemalige der Technischen Universität Chemnitz.

Master of Science / Department of Computing and Information Sciences / William H. Hsu / Entity search has been proposed as a search method for domain-specific Internet applications. It differs from the classical approaches used by search engines which give a "page-view result": listing the URLs of web pages containing the desired keywords. Entity search returns more structured results listing the specific information that a user seeks, such as an email address or a phone number. It not only provides the URL links to targets, but also attributes of target entities (e.g., email address, phone number, etc.). Compared to classical search methods, entity search is a more direct and user-friendly method for searching through a large volume of web documents. After the user submits a query, the extracted entities are ordered by their relevance to the query. While previous work has proposed various complex formulas for entity ranking, it has not been shown whether such complexity is needed. In this research I explore the problem of whether a simpler method can achieve reasonable results. I have designed an entity-search and ranking algorithm using a formula that simply combines a page’s PageRank and an entity's distance to the query keywords to produce a metric for ranking discovered entities. My research goal is to answer the question of whether effective entity ranking can be performed by an algorithm that computes matching scores specific to the entity search domain, and what improvements are necessary to refine the result. My approach takes into account the entity's proximity to the keywords in the query as well as the quality of the page where it is contained. I implemented a system based on the algorithm and perform experiments to show that in most cases the result is consistent with the user's desired outcome.

published_or_final_version / Statistics and Actuarial Science / Master / Master of Philosophy

Each year, admissions officers throughout the United States commit many intense months to reviewing applications to their college/university. According to the College Board, there are established key elements considered in admissions decisions, including grades in college prep courses, standardized test scores, overall academic performance, and class rank. Approximately half of high schools in the U.S. provide class rank, yet it has maintained importance as the number four factor for over a decade, trumping other factors such as extracurricular accomplishments, teacher recommendations, and interviews. A student’s rank-in-class can be used to determine their relative achievement within his or her school, to compare them to the entire applicant pool at a college or university, and to rate students for scholarship selection, along with selections for countless other accolades and financial awards. Rank is calculated across a wide span of methods using grade point averages (GPAs) that sometimes account for course rigor, and sometimes do not. So that colleges/universities might evaluate rigor and competitiveness of each applicant based on the school’s institutional priorities, I contend that colleges/universities should recalculate GPAs as provided from the high school, giving weight to what they value as an institution. Over the past year, I have dramatically shifted my belief in the way rank ought to be used. Earlier in my admissions career, I believed rank was accurate and useful. Now that I have taken significant time to consider the role of rank from the perspective of a school counselor, I realize that it is not the beacon of precision. It has become increasingly clear to me that it is the job of colleges/universities to rank high school students; it is not the job of high schools. During months spent speaking with current and former school counselors, and my own motivation to become a school counselor, I realized that it does not ultimately benefit high schools to provide colleges with rank and it does not benefit colleges to use a precise rank that is born out of one specific context.

The goal of diploma thesis is describing and evaluating of present, officially published rankings in the Czech Republic and creating own rankings in one branch of economy, specifically in health insurance industry. Evaluation is focused on quality and availability of information and used criteria, their significance and evaluators. I also focused on users of rankings and ways of measuring companies' efficiency, which is one of the most important criteria of rankings. Second part is focused on health insurance market. I describe subjects on this market, analyze share of different kinds of insurance on insurance globally and compare to Czech and European economy. I analyze evaluation of insurance companies by Czech national bank, which is authority for inspection of insurance market or rating agencies.

在日常生活和科学研究中产生大量偏好数据，其反应一组被关注对象受偏好的程度。通常用排序数据或多元选择数据来记录观察结果。有时候关于两个对象的偏好没有明显强弱之分，导致排序产生节点，也就是所谓的删失排序。为了研究带有删失的排序数据，基于Thurstone的随机效用假设理论我们建立了一个对称贝叶斯probit模型。然而，参数识别是probit模型必须解决的问题，即确定一组潜在效用的位置和尺度。通常方法是选择其中一个对象为基，然后用其它对象的效用减去这个基的效用，最后我们关于这些效用差来建模。问题是，在用贝叶斯方法处理多元选择数据时，其预测结果对基的选择有敏感性，即选不同对象为基预测结果是不一样的。本文，我们虚构一个基，即一组对象偏好的平均。依靠这个基，我们为多元选择probit模型给出一个不依赖于对象标号的识别方法，即对称识别法。进一步，我们设计一种贝叶斯算法来估计这个模型。通过仿真研究和真实数据分析，我们发现这个贝叶斯probit模型被完全识别，而且消除通常识别法所存在的敏感性。接下来，我们把这个关于多元选择数据建立的probit模型推广到处理一般删失排序数据，即得到对称贝叶斯删失排序probit 模型。最后，我们用这个模型很好的分析了香港赌马数据。 / Vast amount of preference data arise from daily life or scientific research, where observations consist of preferences on a set of available objects. The observations are usually recorded by ranking data or multinomial data. Sometimes, there is not a clear preference between two objects, which will result in ranking data with ties, also called censored rank-ordered data. To study such kind of data, we develop a symmetric Bayesian probit model based on Thurstone's random utility (discriminal process) assumption. However, parameter identification is always an unavoidable problem for probit model, i.e., determining the location and scale of latent utilities. The standard identification method need to specify one of the utilities as a base, and then model the differences of the other utilities subtracted by the base. However, Bayesian predictions have been verified to be sensitive to specification of the base in the case of multinomial data. In this thesis, we set the average of the whole set of utilities as a base which is symmetric to any relabeling of objects. Based on this new base, we propose a symmetric identification approach to fully identify multinomial probit model. Furthermore, we design a Bayesian algorithm to fit that model. By simulation study and real data analysis, we find that this new probit model not only can be identifed well, but also remove sensitivities mentioned above. In what follows, we generalize this probit model to fit general censored rank-ordered data. Correspondingly, we get the symmetric Bayesian censored rank-ordered probit model. At last, we apply this model to analyze Hong Kong horse racing data successfully. / Detailed summary in vernacular field only. / Pan, Maolin. / Thesis (Ph.D.)--Chinese University of Hong Kong, 2013. / Includes bibliographical references (leaves 50-55). / Electronic reproduction. Hong Kong : Chinese University of Hong Kong, [2012] System requirements: Adobe Acrobat Reader. Available via World Wide Web. / Abstract also in Chinese. / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Overview --- p.2 / Chapter 1.1.1 --- The Ranking Model --- p.2 / Chapter 1.1.2 --- Discrete Choice Model --- p.4 / Chapter 1.2 --- Methodology --- p.7 / Chapter 1.2.1 --- Data Augmentation --- p.8 / Chapter 1.2.2 --- Marginal Data Augmentation --- p.8 / Chapter 1.3 --- An Outline --- p.9 / Chapter 2 --- Bayesian Multinomial Probit Model Based On Symmetric I-denti cation --- p.11 / Chapter 2.1 --- Introduction --- p.11 / Chapter 2.2 --- The MNP Model --- p.14 / Chapter 2.3 --- Symmetric Identification and Bayesian Analysis --- p.17 / Chapter 2.3.1 --- Symmetric Identification --- p.18 / Chapter 2.3.2 --- Bayesian Analysis --- p.21 / Chapter 2.4 --- Case Studies --- p.25 / Chapter 2.4.1 --- Simulation Study --- p.25 / Chapter 2.4.2 --- Clothes Detergent Purchases Data --- p.27 / Chapter 2.5 --- Summary --- p.29 / Chapter 3 --- Symmetric Bayesian Censored Rank-Ordered Probit Model --- p.30 / Chapter 3.1 --- Introduction --- p.30 / Chapter 3.2 --- Ranking Model --- p.33 / Chapter 3.2.1 --- Ranking Data --- p.33 / Chapter 3.2.2 --- Censored Rank-Ordered Probit Model --- p.35 / Chapter 3.2.3 --- Symmetrically Identified CROP Model --- p.36 / Chapter 3.3 --- Bayesian Analysis on Symmetrically Identified CROP Model --- p.37 / Chapter 3.3.1 --- Model Estimation --- p.38 / Chapter 3.4 --- Application: Hong Kong Horse Racing --- p.41 / Chapter 3.5 --- Summary --- p.44 / Chapter 4 --- Conclusion and Further Studies --- p.45 / Chapter A --- Prior for covariance matrix with trace augmented restriction --- p.47 / Chapter B --- Derivation of sampling intervals --- p.49 / Bibliography --- p.50

Adverse Drug Reactions (ADRs) are a major cause of morbidity and mortality worldwide. Clinical trials, which are extremely costly, human labor intensive and specific to controlled human subjects, are ineffective to uncover all ADRs related to a drug. There is thus a growing need of computing-supported methods facilitating the automated detection of drugs-related ADRs from large reports data sets; especially ADRs that left undiscovered during clinical trials but later arise due to drug-drug interactions or prolonged usage. For this purpose, big data sets available through drug-surveillance programs and social media provide a wealth of longevity information and thus a huge opportunity. In this research, we thus design a system using machine learning techniques to discover severe unknown ADRs triggered by a combination of drugs, also known as drug-drug-interaction. Our proposed Multi-drug Adverse Reaction Analytics System (MARAS) adopts and adapts an association rule mining-based methodology by incorporating contextual information to detect, highlight and visualize interesting drug combinations that are strongly associated with a set of ADRs. MARAS extracts non-spurious associations that are true representations of the combination of drugs taken and reported by patients. We demonstrate the utility of MARAS via case studies from the medical literature, and the usability of the MARAS system via a user study using real world medical data extracted from the FDA Adverse Event Reporting System (FAERS).