My research is about monitoring the software development process to assess Capability maturity level. Capability Maturity Model (CMM) was developed to improve the software process based on subjective assessment by teams of experts. We propose an objective CMM assessment, which replaces expensive and time-consuming human effort by a knowledge-based system. Compared to Subjective CMM assessment, Objective CMM assessment can be less expensive, takes less time, and is easy to estimate the software development environment maturity. The accuracy of Objective CMM assessment can be the same as Subjective CMM assessment if enough activities are represented as objective activities. For example, if subjective activities total 80 % and objective activities total 20 %, then the accuracy of Objective CMM assessment is not reliable. It would be reliable if the objective activity is increased up to 80% from 20%. This dissertation presents how to change from Subjective CMM assessment to Objective CMM assessment, and we will prove that Objective CMM Assessment is effective.

A retrospective research study was conducted to determine the effectiveness of the Drug Evaluation and Classification Program (DECP) procedures in order to identify subjects under the influence of specific drug categories. The investigator wanted to determine if the procedures are reliable and whether a drug recognition expert (DRE) can properly apply the DECP procedures to consistently identify the drug category and have that opinion supported by toxicology. A total of 324 enforcement drug influence evaluations (DIEs) were obtained from the DRE data tracking system (DRE-DTS) that is maintained through the National Highway Traffic Safety Administration (NHTSA). Toxicology results related to each DIE were compared to the DRE??s drug category prediction. The objectives of the study were to determine if the 12-step DECP process enables DREs to identify drug categories and those most frequently identified and confirmed in Texas. Using the DECP, the DRE??s ability to identify specific drugs according to categories and to have the identification supported by toxicology was moderately accurate at best. Of the 324 evaluations that had toxicology results, the DRE correctly identified drug categories as follows: depressants, 60.5%; stimulants, 32%; hallucinogens, 12%; PCP, 46.6%; narcotic analgesics, 51.6%; inhalants, 14.2%; and cannabis, 64.9%. To determine which drug categories were called most frequently, the enforcement DIEs were analyzed according to each specific category. The investigator found the following drug categories were most frequently called by DREs who evaluated subjects: depressants, 182; cannabis, 142; narcotic analgesics, 83; stimulants, 62; PCP, 18; inhalants, 4; and hallucinogens, 1. To determine which drug categories were most frequently confirmed through toxicology, the enforcement DIEs were analyzed for the number of confirmations for each specific drug category. The investigator found the following drug categories were most frequently confirmed through toxicology: depressants, 176; cannabis, 140; stimulants, 106; narcotic analgesics, 99; PCP, 26; hallucinogens, 8; and inhalants, 4. The investigator was able to determine, based on the DRE-DTS data, that Texas DREs are only moderately accurate in identifying drug categories when utilizing the DECP procedures in enforcement settings. Furthermore, the DRE??s prediction of drug category was not consistently supported by toxicology results obtained from evaluated subjects.

This dissertation has focused on automated power system fault analysis. New contributions to fault section estimation, protection system performance evaluation and power system/protection system interactive simulation have been achieved. Intelligent techniques including expert systems, fuzzy logic and Petri-nets, as well as data from remote terminal units (RTUs) of supervisory control and data acquisition (SCADA) systems, and digital protective relays have been explored and utilized to fufill the objectives. The task of fault section estimation is difficult when multiple faults, failures of protection devices, and false data are involved. A Fuzzy Reasoning Petri-nets approach has been proposed to tackle the complexities. In this approach, the fuzzy reasoning starting from protection system status data and ending with estimation of faulted power system section is formulated by Petri-nets. The reasoning process is implemented by matrix operations. Data from RTUs of SCADA systems and digital protective relays are used as inputs. Experiential tests have shown that the proposed approach is able to perform accurate fault section estimation under complex scenarios. The evaluation of protection system performance involves issues of data acquisition, prediction of expected operations, identification of unexpected operations and diagnosis of the reasons for unexpected operations. An automated protection system performance evaluation application has been developed to accomplish all the tasks. The application automatically retrieves relay files, processes relay file data, and performs rule-based analysis. Forward chaining reasoning is used for prediction of expected protection operation while backward chaining reasoning is used for diagnosis of unexpected protection operations. Lab tests have shown that the developed application has successfully performed relay performance analysis. The challenge of power system/protection system interactive simulation lies in modeling of sophisticated protection systems and interfacing the protection system model and power system network model seamlessly. An approach which utilizes the "compiled foreign model" mechanism of ATP MODELS language is proposed to model multifunctional digital protective relays in C++ language and seamlessly interface them to the power system network model. The developed simulation environment has been successfully used for the studies of fault section estimation and protection system performance evaluation.

Rule induction is a popular technique for knowledge acquisition and data mining. Many techniques, such as ID3, C4.5, CART (tree induction tecniques) and Artificial Neural Networks have been developed and widely used. However, most techniques are either based on categorical or numerical mechanisms to assess the importance of different input variables, which may not produce the optimal rule when a mixture of variables exists. In 1992, Liang proposed a composite approach called CRIS that use different method to analyze different types of data in inducing rules for binary classification. Yang conducted a follow-up research to extend the original algorithm to multiple categories. However, both methods do not take variable interaction into consideration. The purpose of this research is to extend previous approach and extend by including second-order interaction. We also take into consideration the kurtosis and skewness of data for numerical variables. For categorical data, we also adopt ID3 algorithm to handle classes with low representation in the sample. In order to evaluate this technique, we develop a prototype CRIS 3.0 and compare with existing techniques, including multi-category-CRIS, CART and C4.5 as benchmark. The results show that CRIS 3.0 has the highest probability of producing the highest prediction accuracy.

Thesis (M.S.)--West Virginia University, 2004. / Title from document title page. Document formatted into pages; contains xiii, 148 p. : ill. (some col.). Includes abstract. Includes bibliographical references (p. 90-95).

Thesis ( Ph.D. ) -- University of Texas at Arlington, 2008.

Thesis (Ph. D.)--University of Washington, 2003. / Vita. Includes bibliographical references (leaves 223-231).

Thesis (M.S.)--West Virginia University, 2001. / Title from document title page. Document formatted into pages; contains xii, 121 p. : ill. (some col.). Includes abstract. Includes bibliographical references (p. 118-121).

Thesis (M.S.)--Worcester Polytechnic Institute. / Keywords: run-time monitoring; continual validation; software probes; probing. Includes bibliographical references (p. 98-101).

Thesis (Ph. D.)--University of Chicago, Committee on Medical Physics, December 2003. / Includes bibliographical references. Also available on the Internet.