Browsing Theses by Subject "Data quality"
Now showing items 1-2 of 2
Development of Artificial Intelligence-based In-Silico Toxicity Models. Data Quality Analysis and Model Performance Enhancement through Data Generation.Toxic compounds, such as pesticides, are routinely tested against a range of aquatic, avian and mammalian species as part of the registration process. The need for reducing dependence on animal testing has led to an increasing interest in alternative methods such as in silico modelling. The QSAR (Quantitative Structure Activity Relationship)-based models are already in use for predicting physicochemical properties, environmental fate, eco-toxicological effects, and specific biological endpoints for a wide range of chemicals. Data plays an important role in modelling QSARs and also in result analysis for toxicity testing processes. This research addresses number of issues in predictive toxicology. One issue is the problem of data quality. Although large amount of toxicity data is available from online sources, this data may contain some unreliable samples and may be defined as of low quality. Its presentation also might not be consistent throughout different sources and that makes the access, interpretation and comparison of the information difficult. To address this issue we started with detailed investigation and experimental work on DEMETRA data. The DEMETRA datasets have been produced by the EC-funded project DEMETRA. Based on the investigation, experiments and the results obtained, the author identified a number of data quality criteria in order to provide a solution for data evaluation in toxicology domain. An algorithm has also been proposed to assess data quality before modelling. Another issue considered in the thesis was the missing values in datasets for toxicology domain. Least Square Method for a paired dataset and Serial Correlation for single version dataset provided the solution for the problem in two different situations. A procedural algorithm using these two methods has been proposed in order to overcome the problem of missing values. Another issue we paid attention to in this thesis was modelling of multi-class data sets in which the severe imbalance class samples distribution exists. The imbalanced data affect the performance of classifiers during the classification process. We have shown that as long as we understand how class members are constructed in dimensional space in each cluster we can reform the distribution and provide more knowledge domain for the classifier.
Measurement properties of respondent-defined rating-scales. An investigation of individual characteristics and respondent choices.It is critical for researchers to be confident of the quality of survey data. Problems with data quality often relate to measurement method design, through choices made by researchers in their creation of standardised measurement instruments. This is known to affect the way respondents interpret and respond to these instruments, and can result in substantial measurement error. Current methods for removing measurement error are post-hoc and have been shown to be problematic. This research proposes that innovations can be made through the creation of measurement methods that take respondents¿ individual cognitions into consideration, to reduce measurement error in survey data. Specifically, the aim of the study was to develop and test a measurement instrument capable of having respondents individualise their own rating-scales. A mixed methodology was employed. The qualitative phase provided insights that led to the development of the Individualised Rating-Scale Procedure (IRSP). This electronic measurement method was then tested in a large multi-group experimental study, where its measurement properties were compared to those of Likert-Type Rating-Scales (LTRSs). The survey included pre-validated psychometric constructs which provided a baseline for comparing the methods, as well as to explore whether certain individual characteristics are linked to respondent choices. Structural equation modelling was used to analyse the survey data. Whilst no strong associations were found between individual characteristics and respondent choices, the results demonstrated that the IRSP is reliable and valid. This study has produced a dynamic measurement instrument that accommodates individual-level differences, not addressed by typical fixed rating-scales.