Search

This dissertation describes research into “Antennas with Frequency Domain Control for Future Communication Systems” and several novel antennas are shown, each of which addresses a specific issue for current and future communication systems, in terms of wideband coverage, channel capacity, antenna isolation and band-rejection. These antenna designs may be candidates for implementation in future multiband radios, and software defined radio (SDR) and cognitive radio (CR) systems, which are two new concepts in wireless communications in the foreseeable future, although it is evident that there are as yet no clear specifications for those future systems. A novel two-port reconfigurable antenna which can operate within a narrowband or wideband mode is presented. Three different structures of wideband reconfigurable balanced antennas, with a wide tuning range, have been proposed. When the balanced antenna is combined with the two-port chassis antenna, it becomes a reconfigurable MIMO antenna for small terminals and at least 15 dB of isolation is achieved. Several designs of conical monopole antennas, incorporating different types of slots to achieve good band-rejection behaviour, have been introduced. These are the 2 Cshaped, 4 C-shaped slots, 4 U-shaped slots, 4 tilted-U-shaped slots and 4 U-C-shaped slots. The study of wideband antennas with notched-band behaviour using a simple equivalent circuit model has been proposed. It has been noted that increasing the number of resonators and the coupling factor will increase the band-rejection. However, it will also widen the bandwidth of the frequency notched band. A novel pyramidal monopole antenna, with four loop shaped slots, offering wide tunable band-notch, is also presented.

The Detection of surgical adverse events has become increasingly important with the growing demand for quality improvement and public health surveillance with surgery. Event reporting is one of the key steps in determining the impact of postoperative complications from a variety of perspectives and is an integral component of improving transparency around surgical care and ultimately around addressing complications. Manual chart review is the most commonly used method in identification of adverse events. Though the manual chart review is the most commonly used method that is considered the "gold-standard" for detecting adverse events for many patient safety studies (research setting), it could be very labor-intensive and time-consuming and thus many hospitals have found it too expensive to routinely use. In this dissertation, aiming to accelerate the process of extracting postoperative outcomes from medical charts, an automated postoperative adverse events detection application has been developed by using structured electronic health record (EHR) data and unstructured clinical notes. First, pilot studies are conducted to test the feasibility by using only completed EHR data and focusing on three types of surgical site infection (SSI). The built models have high specificity as well as very high negative predictive values, reliably eliminating the vast majority of patients without SSI, thereby significantly reducing the chart reviewers' burden. Practical missing data treatments have also been explored and compared. To address modeling challenges, such as high-dimensional dataset, and imbalanced distribution, several machine learning methods haven been applied. Particularly, one single-task and five multi-task learning methods are developed and compared for their detection performance. The models demonstrated high detection performance, which ensures the feasibility of accelerating the manual process of extracting postoperative outcomes from medical chart. Finally, the use of structured EHR data, clinical notes and the combination of these data types have been separately investigated. Models using different types of data were compared on their detection performance. Models developed with very high AUC score have demonstrated that supervised machine learning methods can be effective for automated detection of surgical adverse events. [The dissertation citations contained here are published with the permission of ProQuest LLC. Further reproduction is prohibited without permission. Copies of dissertations may be obtained by Telephone (800) 1-800-521-0600. Web page: http://www.proquest.com/en-US/products/dissertations/individuals.shtml.]

The Detection of surgical adverse events has become increasingly important with the growing demand for quality improvement and public health surveillance with surgery. Event reporting is one of the key steps in determining the impact of postoperative complications from a variety of perspectives and is an integral component of improving transparency around surgical care and ultimately around addressing complications. Manual chart review is the most commonly used method in identification of adverse events. Though the manual chart review is the most commonly used method that is considered the “gold-standard” for detecting adverse events for many patient safety studies (research setting), it could be very labor-intensive and time-consuming and thus many hospitals have found it too expensive to routinely use. In this dissertation, aiming to accelerate the process of extracting postoperative outcomes from medical charts, an automated postoperative adverse events detection application has been developed by using structured electronic health record (EHR) data and unstructured clinical notes. First, pilot studies are conducted to test the feasibility by using only completed EHR data and focusing on three types of surgical site infection (SSI). The built models have high specificity as well as very high negative predictive values, reliably eliminating the vast majority of patients without SSI, thereby significantly reducing the chart reviewers’ burden. Practical missing data treatments have also been explored and compared. To address modeling challenges, such as high-dimensional dataset, and imbalanced distribution, several machine learning methods haven been applied. Particularly, one single-task and five multi-task learning methods are developed and compared for their detection performance. The models demonstrated high detection performance, which ensures the feasibility of accelerating the manual process of extracting postoperative outcomes from medical chart. Finally, the use of structured EHR data, clinical notes and the combination of these data types have been separately investigated. Models using different types of data were compared on their detection performance. Models developed with very high AUC score have demonstrated that supervised machine learning methods can be effective for automated detection of surgical adverse events.

Software repackaging is a common approach for creating malware. In this approach, malware authors inject malicious payloads into legitimate applications; then, to ren- der security analysis more difficult, they obfuscate most or all of the code. This forces analysts to spend a large amount of effort filtering out benign obfuscated methods in order to locate potentially malicious methods for further analysis. If an effective mechanism for filtering out benign obfuscated methods were available, the number of methods that must be analyzed could be reduced, allowing analysts to be more productive. In this thesis, we introduce SEMEO, a highly effective and efficient fil- tering approach that can determine whether an obfuscated and an original version of a method are semantically equivalent. Our approach handles seven common, com- plex types of obfuscation and can be effective even when all types are compositely applied. In an empirical evaluation, we applied SEMEO to nine Android apps of varying complexity, and the approach provided over 76% recall and 100% precision in identifying semantically equivalent methods. We then performed three additional studies, that showed that: (1) SEMEO is much more effective at identifying semantically equivalent methods than FSquaDRA, an existing technique; (2) SEMEO is also effective for identifying repackaged apps that have been previously obfuscated by ProGuard, a popular obfuscation tool; and (3) SEMEO is effective at identifying semantically equivalent methods in a repackaged, malicious version of Pokemon Go.

A survival response study was carried out by using D. melanogaster and the opportunistic pathogen B. cereus as the agent of selection. The spores of B. cereus, a gram-positive bacteria that can cause the human pathogen disease, were applied in our artificial laboratory selection. Selected lines were treated with B. cereus spores. Wound control lines were punctured with a needle dipped into sterile H2O. Control lines did not apply any treatment. Three different environmental treatments were used within each line type (autoclaved spores of B.cereus, sterile H2O and no treatment). The autoclaved spores were used as an inducer of the immune response in our study, with the purpose of boosting their innate immunity. Our hypothesis was that selected lines will live longer as they are immune to “B. cereus” spores. By comparing the average mortality rate of different line types, flies in selected lines were observed to die slowest, which was correspondent to our hypothesis. Selected and control lines were studied in different stress environment (starvation, desiccation, chill coma, oxidation) to investigate their survival response to stress. Those flies resistant to B. cereus spores in our selection were expected to survive longer than the flies in other line type under different stresses. To study the trade-off for these survival responses, energy components were measured. The result showed that there was a change in lipid and glycogen concentration in selected flies, which confirmed our trade-off hypothesis between survival responses and energy component. Advisor: Lawrence G. Harshman

As other programs in education, international collaborative programs face challenges of quality assurance and accountability for students' learning outcomes. Through a survey of 1458 students, the present research employs a "program logic model" (input-activity-outcome) to measure students' perceived global learning outcomes in international collaborative programs in Shanghai, China. Multiple regressions were used to test the relationship between different aspects of students' learning experience and their perceived learning outcomes in the programs. Findings of this research suggest that students' learning experience had positive influence on students' perception of learning outcomes, even when students' personal variables, such as gender, grade level, major, pre-college grades, and international experience were controlled. This research further investigated the possible moderating effect of personal variables on the relationship between learning experience and perceived learning outcomes. Findings from the analysis show that the impact of students' learning experience on perceived learning outcomes may vary by pre-college grades and prior international experience. Although these differences are small in magnitude, they suggest that the effects of students' learning experience on their perceived learning outcomes should be carefully examined.

Clustering is an important research problem for knowledge discovery from databases. It focuses on finding hidden structures embedded in datasets. It is non-trivial to arrive at a clustering in a dataset such that each pair of data points within the same cluster is similar to each other, and each pair in different clusters is distinct from each other. This is due to the multiplicity of meanings of similarity between data points and also from criteria determining the number, shape, and boundaries of clusters. Despite a large body of published research, new clustering problems keep arising requiring novel solutions. Such a situation is evolving in the field of biomedical research which is generating a large number of interrelated and interdependent datasets, and also in many other domains of science and business. We have developed three novel methodologies for clustering to meet these newly emerging needs. The first problem we have solved relates to the grouping of data points with “similar” density in the data space into distinct clusters, using full dimensional clustering. Based on the pair-wise similarity matrix among data points, we define a new type of relationship among them - that of the point pairs being Mutual K-Nearest Neighbors (MKNN) of each other, and design clustering algorithms based on this new notion to capture the data density. Compared with traditional Euclidean distance based clustering algorithms for datasets having different densities, our MKNN-based clustering algorithm allows users to form density-based clusters with significantly lower sensitivity to parameters . We have analytically and empirically demonstrated, using both synthetic and real-world datasets, the increased capability, precision, efficiency, and robustness of our algorithm. The second clustering algorithm which we have developed incorporates prior domain knowledge, provided as pair-wise similarity matrix in one dataset, into the clustering performed for data in another dataset. The data objects in “prior knowledge” data source and the second data source are the same. By adopting a semi-supervised clustering procedure, our algorithm, called Semi-supervised Gaussian Infinite Mixture Model (SGIMM), balances information from two data sources and generates clusters enforcing precise pair-wise relationships. SGIMM accommodates many types of prior knowledge and from the empirical studies done with both the synthetic data and the real-world data; SGIMM generates high quality clusters regardless of the quality of prior knowledge. The third type of problem we have solved relates to the discovery of subspace clusters. Numerous real world applications focus on selecting subsets of data points and feature subspaces having desirable characteristics specified in terms of properties such as low variance, high distinction, low residue value, etc. We use lattice structured search spaces to identify low variance subspace clusters from one dataset (bicluster), two datasets (3-Cluster), and high discrepancy subspace clusters from a single dataset (polarized bicluster). The results on both synthetic datasets and genomic datasets have been shown for all these types of clustering tasks and they show performance better than what is shown by most of the existing algorithms.