11,874 Matching Results

Search Results

Co-training and Self-training for Word Sense Disambiguation

Description: This paper investigates the application of co-training and self-training to word sense disambiguation. Optimal and empirical parameter selection methods for co-training and self-training are investigated, with various degrees of error reduction. A new method that combines co-training with majority voting is introduced, with the effect of smoothing the bootstrapping learning curves, and improving the average performance.
Date: May 2004
Creator: Mihalcea, Rada, 1974-
Partner: UNT College of Engineering

Finding Semantic Associations on Express Lane

Description: This paper introduces a new codification scheme for efficient computation of measures in semantic networks. The scheme is particularly useful for fast computation of semantic associations between words and implementation of an informational retrieval operator for efficient search in semantic spaces. Other applications may also be possible.
Date: May 2004
Creator: Nastase, Vivi & Mihalcea, Rada, 1974-
Partner: UNT College of Engineering

Graph-based Ranking Algorithms for Sentence Extraction, Applied to Text Summarization

Description: Abstract: This paper presents an innovative unsupervised method for automatic sentence extraction using graph-based ranking algorithms. We evaluate the method in the context of a text summarization task, and show that the results obtained compare favorably with previously published results on established benchmarks.
Date: July 2004
Creator: Mihalcea, Rada, 1974-
Partner: UNT College of Engineering

SenseLearner: Minimally Supervised Word Sense Disambiguation for All Words in Open Text

Description: This paper introduces SenseLearner - a minimally supervised sense tagger that attempts to disambiguate all content words in a text using the sense from WordNet. SenseLearner participated in the SENSEVAL-3 English all words task, and achieved an average accuracy of 64.6%.
Date: 2004
Creator: Mihalcea, Rada, 1974- & Faruque, Ehsanul
Partner: UNT College of Engineering

TextRank: Bringing Order into Texts

Description: In this paper, the authors introduce TextRank, a graph-based ranking model for text processing, and show how this model can be successfully used in natural language applications.
Date: July 2004
Creator: Mihalcea, Rada, 1974- & Tarau, Paul
Partner: UNT College of Engineering

The SENSEVAL-3 English Lexical Sample Task

Description: This paper presents the task definition, resources, participating systems, and comparative results for the English lexical sample task, which was organized as part of the SENSEVAL-3 evaluation exercise.
Date: July 2004
Creator: Mihalcea, Rada, 1974-; Chklovski, Timothy A. (Timothy Anatolievich), 1977- & Kilgarriff, Adam
Partner: UNT College of Engineering

Social skills training for individuals with schizophrenia: Evaluation of treatment outcome and acquisition of social and cognitive skills.

Description: Social and cognitive skill acquisition were evaluated in 33 (male=24, female=11) outpatients with schizophrenia or schizoaffective disorder. A social skills training treatment group (n=19) was compared to a wait-list control (n=14). Participants' mean age was 41 years, mean number of hospitalizations 10.4, and mean number of years with diagnosis 15.8. Assessment measures included WAIS-III Picture Arrangement subtest, Social Cue Recognition Test, COGLAB, WMS-III Word List subtest, and SADS-C. Results did not support the main hypotheses of improved social and cognitive skills in the treatment group. Participants with better memory and attention at pre-testing also did not show an advantage in social skills improvement. Contrary to hypotheses, the control group improved the most on some social and cognitive measures. Several supplemental hypotheses yielded the following results: lack of volunteer participation from paranoid schizophrenia individuals; evidence that schizoaffective disorder participants may be less cognitively impaired and better able to benefit from social skills training; and younger, less chronic participants with better attentional capacities may benefit most from social skills training. Findings are discussed in light of the possibility that improving social skills might not improve social and cognitive functioning, at least with the dosage of social skills training provided in this study. Limitations such as a sampling bias and small study size are also considered as possible explanations for the pattern of findings. Clinical and research implications are discussed to apply and extend the current findings.
Date: December 2004
Creator: Conner, Dianna Holden
Partner: UNT Libraries

Improving Performance in a Global Logistics Company: Operational Performance Before and After Process Improvement

Description: The purpose of this study was to determine the effectiveness of an intervention designed to eliminate damage notification failures in a customer-specific standard operating procedure used by a global logistics company. Process maps identified locations in the process where damage notification failures could most likely occur. A revised process was designed overnight to eliminate as many notification failure points as possible. In addition, a job aid was included to help facilitate the process change for the drivers. The results of the intervention showed a rapid and profound decrease in damage notification failures leading to the retention of a large, profitable account with a minimal initial investment of time and money.
Date: December 2004
Creator: Dearman, Shawn Kale
Partner: UNT Libraries

Student Outcomes in Selected Distance Learning and Traditional Courses for the Dallas County Community College District: A Pilot Study

Description: The study compared outcomes for distance learning courses with those of traditional courses offered by the seven campuses of Dallas County Community College District (DCCCD). The course outcomes were defined as completion rate, dropout rate and success rate. Eleven courses offered during the fall 2003 semester were selected for the study. The methods of instruction employed for each course were traditional classroom lecture/discussion and distance learning formats of Internet, TeleCourse and TeleCourse Plus. Internet courses are delivered on-line, using Internet access and a browser, TeleCourse uses one-way videos or public broadcasting, and TeleCourse Plus is a hybrid between Internet and TeleCourse courses. Seven of the courses selected were part of the core curriculum approved by Texas Higher Education Coordinating Board (THECB) while four other courses were completely transferable. Two types of specific data were extracted: course data and individual student data. Course data included method of instruction, length of course, instructor's load, enrollment, number of withdrawals, and grade distribution. In addition, course requirements including the use of email, videos and Internet, orientation and testing on campus were added as variables. The student data included demographic variables such as age, gender, ethnicity, family status, employment and academic variables including number of credit hours completed, previous distance learning courses, grade point average (GPA), grades, placement scores, previous degrees held, withdrawal history, and financial aid. The theoretical framework for ensuring sound statistical analysis was Astin's student engagement model. The results showed that significant differences exist due to the three distance learning methods of instruction for all course outcomes studied. Completion and success rates are higher for traditional courses and dropout rate is higher for distance learning ones. The outcomes for Internet courses are closer to the rates of traditional courses. Student factors that relate to performance in distance learning courses are GPA, credit ...
Date: December 2004
Creator: Borcoman, Gabriela
Partner: UNT Libraries