Exploring the integration of data mining and data visualization

Due to the rapid advances in computing and sensing technologies, enormous amounts of data are being generated everyday in various applications. The integration of data mining and data visualization has been widely used to analyze these massive and complex data sets to discover hidden patterns. For both data mining and visualization to be effective, it is important to include the visualization techniques in the mining process and to generate the discovered patterns for a more comprehensive visual view. In this dissertation, four related problems: dimensionality reduction for visualizing high dimensional datasets, visualization-based clustering evaluation, interactive document mining, and multiple clusterings exploration are studied to explore the integration of data mining and data visualization. In particular, we 1) propose an efficient feature selection method (reliefF + mRMR) for preprocessing high dimensional datasets; 2) present DClusterE to integrate cluster validation with user interaction and provide rich visualization tools for users to examine document clustering results from multiple perspectives; 3) design two interactive document summarization systems to involve users efforts and generate customized summaries from 2D sentence layouts; and 4) propose a new framework which organizes the different input clusterings into a hierarchical tree structure and allows for interactive exploration of multiple clustering solutions.

An investigation into the comparable effectiveness of two tests of writing administered to college-level English as a Second Language students: The Computerized Placement Test and holistically scored essay exams

This dissertation describes the findings and implications of a correlational analysis. Scores earned on the Computerized Placement Test (CPT), sentence skills, were compared to essay scores of advanced English as a Second Language (ESL) students. As the CPT is designed for native speakers of English, it was hypothesized that it could be an invalid or unreliable instrument for non-native speakers. Florida community college students are mandated to take the CPT to determine preparedness, as are students at many other U.S. and Canadian colleges. If incoming students score low on the CPT, they may be required to take up to three semesters of remedial coursework. It is essential that scores earned by non-native speakers of English accurately reflect their ability level. They constitute a large and growing body of non-traditional students enrolled at community colleges.^ The study was conducted at Miami-Dade Community College, Wolfson Campus, fall 1997. Participants included 106 advanced ESL students who took both the CPT sentence skills test and wrote final essay exams. The essay exams were holistically scored by trained readers. Also, the participants took the Placement Articulation Software Service (PASS) exam, an alternative form of the CPT. Scores on the CPT and essays were compared by means of a Pearson product-moment correlation to validate the CPT. Scores on the CPT and the PASS exam were compared in the same manner to verify reliability. A percentage of appropriate placements was determined by comparing essay scores to CPT cutoff score ranges. Finally, the instruments were evaluated by means of independent-samples t-tests for performance differences between gender, age, and first language groups.^ The results indicate that the CPT sentence skills test is a valid and reliable placement instrument for advanced- level ESL students who intend to pursue community college degrees. The correlations demonstrated a substantial relationship between CPT and essay scores and a marked relationship between CPT and PASS scores. Appropriate placements were made in 86% of the cases. Furthermore, the CPT was found to discriminate equally among the gender, age, and first language groups included in this study. ^

The effect of different instructional groupings on student achievement and attitudes from a computer-based interactive videodisk biology unit

The purpose of this research study was to determine the effect of two different instructional groupings (cooperative and traditional whole-class) on student achievement and attitudes using a computer-based interactive videodisc biology unit. The subjects were 64 high school biology students assigned to two heterogeneous experimental groups, randomly selected from two preassigned summer school biology classes, one honors, the other regular. A two-group, posttest-only, control group experimental research design was utilized. Achievement at three cognitive levels and attitudes towards science laserdisc instruction were measured at the conclusion of the study. The cooperative group consistently outperformed the traditional group in achievement posttest scores. Factorial ANOVA on total (overall) achievement scores indicated that subjects in cooperative groups significantly outperformed those in the traditional group, and also that the instructional group, class level, and gender interacted in an ordinal fashion to make a significant difference in how female and male subjects were affected by the treatments depending on their class (aptitude) level. Regular level females and honors level males performed much better when in cooperative groups, whereas group membership did not appear to make a difference for either honors level females or regular level males. A t-test comparing honors level males revealed that cooperative groups were close to being significantly better in total achievement posttest scores than their traditional group counterparts. Factorial MANOVA comparing the instructional groups at three cognitive levels found no significant difference. Analysis on the attitudes posttest data also revealed that subjects in cooperative groups demonstrated more positive attitudes towards science laserdisc instruction; however these differences were not found to be significant. Significant interactions in attitudes of females and males from different class levels had the opposite effect as achievement: honors level females and regular level males demonstrated more positive attitudes towards science laserdisc instruction when in cooperative groups, whereas group membership did not appear to make a difference for honors level males, and regular level females demonstrated the lowest attitudes ratings of any group when involved in cooperative groups. This contrast between achievement and attitudinal results suggests cross-gender interaction in traditionally defined gender roles.