F. De W. Bolman's Thesis--Columbia University. / Published also without thesis note. Vita. Photocopy (positive) made by University Microfilms. Bibliography: p. [241]-243.

published_or_final_version / Comparative Literature / Doctoral / Doctor of Philosophy

Thesis--Munich. / Includes bibliographical references (p. 90-93).

Functional Dependencies (FDs) are commonly used in data cleaning to identify dirty and inconsistent data values. However, many errors require user input for specific do- main knowledge. For example, let us consider the drugs, Advil and Crocin. FDs will consider these two drugs different because they are not syntactically equal. However, Advil and Crocin are synonyms as they are two different drugs with similar chemical compounds but marketed under distinct names in different countries. While FDs have traditionally been used in existing data cleaning solutions to model syntactic equivalence, they are not able to model broader relationships (e.g., synonym, Is-A (Inheritance)) defined by ontologies. In this thesis, we take a first step to discover a new dependency called Ontology Functional Dependencies (OFDs). OFDs model attribute relationships based on re- lationships in a given ontology. We present two effective algorithms to discover OFDs using synonyms and inheritance relationships. Our discovery algorithms search for minimal OFDs and prune the redundant ones. Both algorithms traverse the search lattice in a level-wise Breadth First Search (BFS) manner. In addition, we have devel- oped a set of pruning rules so that we can avoid considering unnecessary candidates in the search lattice. We present an experimental study describing the performance ivand scalability of our techniques. Experimental results show that both algorithms are effective in practice and discover OFDs efficiently for large datasets with millions of tuples. We also present a qualitative study showing that the discovered OFDs are meaningful with high precision and recall. / Thesis / Master of Science (MSc)

These three epigraphs suggest themes for the major divisions ofthis work (dialectics, the sensuous, and the poetic) and also function to foreground the ontological component within the broader tradition of existential phenomenology in the 20th century. Specifically I want to reflect upon the urgency of Merleau-Ponty’s philosophical considerations of what he called carnal and brute ontology, new concepts which, it seems to me, warrant closer scrutiny and continue to provide deeper, more profound alternatives for contemporary critical philosophies. Merleau-Ponty’s attempt to reconfigure the exhausted philosophical formulation ofsubject-object-meaning was cut short by his untimely death, leaving unfinished his critical work, The Visible and the Invisible, which described anew an ethics borne upon key notions of divergence, reversibility, alterity and (the) flesh. / Thesis / Master of Arts (MA)

No description available.

No description available.

Nowadays, the comparison of educational courses and modules is performed manually by experts in the field of education. The main objective of this research work is to create an approach for the automation of this process. The main contribution of this work is a novel, ontology alignment-based methodology for the automated comparison of academic courses and modules, belonging to the cognitive learning domain. The results of this work are appropriate for such tasks as prior learning and degree recognition, the introduction of joint educational programmes and quality assurance in higher education institutions. The set-theoretical models of an educational course, its modules, learning outcomes and keywords were created and converted to the ontology. The choice of the information to be presented in the ontology was based on the careful analysis of programme specifications, module templates and Bologna recommendations for the comparison of educational courses. Ontology was chosen as the data model due to its ability to formally specify semantics, to represent taxonomies and to make inferences regarding data. The formal grammars of a keyword and a learning outcome were created to enable the semi-automated population of the ontology from the module templates. The corresponding annotators were designed in the General Architecture for Text Engineering 6.1. The algorithm for the comparison of educational courses and modules was based on the alignment of ontologies of their keywords and learning outcomes. A novel measure for calculating the similarity between the action verbs in the learning outcomes was introduced and was utilised. Both the measure and the algorithm were implemented in Java. For evaluation purposes, we utilised the module templates from the De Montfort and the Bauman Moscow State Technical Universities. The automatically produced annotations of the keywords and the learning outcomes were evaluated against a manually created gold standard. The high values of the precision, recall and f-measure proved their quality and their suitability for the task. The results produced by the alignment algorithm were compared with those produced by human judgement. The results returned by the experts and the algorithm were comparable, thus showing that the proposed approach is applicable for the partial automation of the comparison of educational modules and courses.

隨著網際網路的發展，人類能夠獲取的資訊也隨之增加。因此，如何增加獲取資訊的效率便成為重要的研究之一。自動摘要系統的目的即在於協助使用者有效率的閱讀。其根本的問題為：如何從文章中找出重要的資訊並呈現給使用者。本文採用三種方法進行摘要，應用於英文的新聞文件：第一種為利用ontology建立一領域的文章可能的主題資訊，並利用該資訊選出重要的段落作為摘要。第二種方法為建立一領域的ontology後，利用ontology所定義的標籤建構摘要的樣板，再利用該樣版搜尋所需的資訊，並將摘要輸出。第三種方法為利用各種不同的特徵找出文章中較重要的段落。另外，我們也將文章依其主題予以分類，利用不同主題的文章呈現的特徵表現不同，改良原本的特徵選取摘要的方法。本文呈現三種以不同方式獲取文章的主題資訊的方法，並利用該資訊呈現文章中較重要的訊息，經過實驗的評估皆獲得一定的成果。 / In the past decade, the explosively growing number of online articles has made efficient information gathering a challenging necessity. We need ways to absorb the information contained in the news articles effectively. Automatically providing summaries of articles is one way to save people time. The essential problems of automatic summarization are: how to identify the useful information and how to present the results to readers. I compare and analyze three kinds of summarization methods. The first method constructs a domain-depend knowledge based on the ontology approach, then use the ontology for gathering the main topics, and chooses the desired proportion of paragraphs as the summary by gathered topical information. The second method is similar to that taken by information-extraction systems. I organize semantic tags into an ontological structure, and the summarization system learns tags patterns for creating summaries from tagged data. The summarization system creates summaries by extracting useful information from a news article, and replaces the semantic tags with extracts in selected tag patterns. The third method analyzes the effects of several previously proposed features for summarization under different situations. The most important observation is that the effectiveness if these features depends on the topics of the news articles Hence, I collect statistical information about the features for different possible new topics, and apply such conditional probabilistic information for extracting summaries. Effectiveness of these proposed methods vary from case to case, but is believed to be satisfactory based on the experimental results.

No description available.