The classification performance of ensemble decision tree classifiers: a case study of detecting fraud in credit card transactions
- Authors: Chogugudza, Mcdonald
- Date: 2022-11
- Subjects: fraud , Commercial fraud , Accounting fraud
- Language: English
- Type: Master's theses , text
- Identifier: http://hdl.handle.net/10353/27590 , vital:69317
- Description: In this dissertation, we propose ensemble decision tree classifiers as an ideal classification technique for solving the problem of fraud in the domain of credit card transactions. Ensemble tree classifiers have been applied in many areas like speech recognition, image recognition and medical diagnostics and have shown excellent results. At the centre of fraud, credit card fraud has been a major concern. The rise in credit card fraud is largely attributed to the nature in which it can be done. A fraudster does not need to always be physically present to commit fraud making it the number one target for criminals. Card-Not-Present refers to this type of fraud where an electronic transaction can be conducted without the need for a client to be present. This can be done via telephonic calls or the web. To be able to come up with better classifiers it was important for the researcher to first investigate what causes misclassifications in fraud detection systems. A systematic literature review was done to uncover the factors that have been identified as causes of misclassifications. It was discovered that many factors lead to misclassifications and several authors have proposed techniques to handle these factors. However, there is no universal techniques for addressing factors that lead to misclassifications as different domains have different datasets which require different techniques. This study investigates how parameters involved in modelling fraud detection systems impact the classification performance of ensemble decision tree classifiers. The factors that were investigated include sample size, sampling technique, learning method and choice of split criterion and how they affect classification performance. A series of experiments were conducted to investigate how the aforementioned factors contributed to better classifiers. Ecommerce data from Vesta corporation made available on Kaggle was used in the experiments. The data was split into two sets, one for training the models and the other for testing the performance of the models. Accuracy, confusion matrix, precision and recall were used as performance measures. Our results showed that a larger sample size resulted in better classifiers. This is attributed to models having more instances to learn from which covers most patterns of fraudulent transactions. The sampling technique was shown to be pivotal in classification performance as under sampling showed a great reduction in performance as it achieved a maximum accuracy of 89.6223 while oversampling produced increased performance with maximum accuracy of 99.9531. Furthermore, our results showed that the choice of split criterion impacts the performance of ensemble tree classifiers. The use of entropy as the choice of split criterion resulted in better classifiers compared to the use of the Gini index. However, the downside is that entropy requires more time to execute compared to the Gini index. Lastly, the learning method proved to impact the performance of ensemble classifiers. Models that used supervised learning had better performance compared to those that use unsupervised learning in detecting credit card fraud. The conclusions from this research are insightful when designing fraud detection systems that use ensemble decision tree classifiers as base learners. , Thesis (Msci) -- Faculty of Science and Agriculture, 2022
- Full Text:
- Date Issued: 2022-11
- Authors: Chogugudza, Mcdonald
- Date: 2022-11
- Subjects: fraud , Commercial fraud , Accounting fraud
- Language: English
- Type: Master's theses , text
- Identifier: http://hdl.handle.net/10353/27590 , vital:69317
- Description: In this dissertation, we propose ensemble decision tree classifiers as an ideal classification technique for solving the problem of fraud in the domain of credit card transactions. Ensemble tree classifiers have been applied in many areas like speech recognition, image recognition and medical diagnostics and have shown excellent results. At the centre of fraud, credit card fraud has been a major concern. The rise in credit card fraud is largely attributed to the nature in which it can be done. A fraudster does not need to always be physically present to commit fraud making it the number one target for criminals. Card-Not-Present refers to this type of fraud where an electronic transaction can be conducted without the need for a client to be present. This can be done via telephonic calls or the web. To be able to come up with better classifiers it was important for the researcher to first investigate what causes misclassifications in fraud detection systems. A systematic literature review was done to uncover the factors that have been identified as causes of misclassifications. It was discovered that many factors lead to misclassifications and several authors have proposed techniques to handle these factors. However, there is no universal techniques for addressing factors that lead to misclassifications as different domains have different datasets which require different techniques. This study investigates how parameters involved in modelling fraud detection systems impact the classification performance of ensemble decision tree classifiers. The factors that were investigated include sample size, sampling technique, learning method and choice of split criterion and how they affect classification performance. A series of experiments were conducted to investigate how the aforementioned factors contributed to better classifiers. Ecommerce data from Vesta corporation made available on Kaggle was used in the experiments. The data was split into two sets, one for training the models and the other for testing the performance of the models. Accuracy, confusion matrix, precision and recall were used as performance measures. Our results showed that a larger sample size resulted in better classifiers. This is attributed to models having more instances to learn from which covers most patterns of fraudulent transactions. The sampling technique was shown to be pivotal in classification performance as under sampling showed a great reduction in performance as it achieved a maximum accuracy of 89.6223 while oversampling produced increased performance with maximum accuracy of 99.9531. Furthermore, our results showed that the choice of split criterion impacts the performance of ensemble tree classifiers. The use of entropy as the choice of split criterion resulted in better classifiers compared to the use of the Gini index. However, the downside is that entropy requires more time to execute compared to the Gini index. Lastly, the learning method proved to impact the performance of ensemble classifiers. Models that used supervised learning had better performance compared to those that use unsupervised learning in detecting credit card fraud. The conclusions from this research are insightful when designing fraud detection systems that use ensemble decision tree classifiers as base learners. , Thesis (Msci) -- Faculty of Science and Agriculture, 2022
- Full Text:
- Date Issued: 2022-11
The effects of school violence on teaching and learning in selected schools in Amatole West District, Eastern Cape
- Poswa, S
- Authors: Poswa, S
- Date: 2022-11
- Subjects: School violence , Violence in children
- Language: English
- Type: Master's theses , text
- Identifier: http://hdl.handle.net/10353/26075 , vital:64788
- Description: The purpose of this study was to look at the effects of school violence on teaching and learning in three selected schools in the Amathole West education district. The objectives of the study guided the entire project and the theory of socio-ecology by Benbenishty and Astor (2005) underpinned and guided the study. Different forms of school violence, causes of school violence, its effects and measures to curb its spread were discussed. The study made use of the pragmatic paradigm which allowed the researcher to make use of both quantitative and qualitative approaches. The data collection instruments consisted of questionnaires and semi-structured interviews. Three high schools were selected purposively with a sample of 75 respondents. Three principals, five teachers (school management team) and 7 learners (school representative council) participated in the semi-structured interview and 60 learners were randomly selected to take part in the questionnaire completion. Data analysis included coding of questionnaires and running the codes on the SPSS package. There after, results were presented using tables and descriptive analysis. Percentages and numeric figures of different forms of school violence were analysed. Data collected from the semi-interviews was coded according to different themes and categories and presented in form of discussions. Thematic analysis was applied in this regard. Therefore, the themes that emerged from the findings of the study were linked to the reviewed literature and research questions. The researcher cross-checked with the participants whether their responses were captured correctly before drawing any conclusions. The major findings of the study revealed that school violence is very much prevalent in the three selected schools. However, the severity of the problem varies from one school to another. Learners indicated that they experience bullying, verbal abuse, corporal punishment, cyberbullying, theft of property, gang-related violence and sexual harassment. Bullying, verbal abuse, cyberbullying and corporal punishment were rated as the most common and frequent forms of violence. In addition to the findings above, some causes of school violence noted by the respondents included peer pressure, poor parental care, exposure to violence at home or communities, unguided exposure to media, easy access to drugs, alcohol and weapons and teacher absenteeism. Effects of school violence included loss of concentration in class, poor academic results, class disruptions and depression or suicidal thoughts. All participants were allowed to say what they think should be done to curb the spread of school violence. Reasons given included that of ensuring that other departments are included in the committees that deal directly with school violence cases, strict and consistent enforcement of the school code of conduct and the department’s policies on school violence. Most learners and teachers indicated that parents are not present in the lives of their children. Thus, parental involvement was strongly recommended by most participants. The researcher had some recommendations to make. It was recommended that communities should take an active part in fighting drug and alcohol abuse by learners in communities. This could be done by reporting people who sell these drugs and alcohol to persons who are under the age of 18. It was further recommended that the department of education should understand and take advantage of technology. School violence apps can be developed specifically for teenagers. These will ensure that knowledge about school violence is spread to most learners. Lastly, workshops should be conducted with both teachers and learners to equip them more about school violence. The conclusion of the study highlighted some important points to be taken into consideration. One of the points was that a failure to implement the strategies outlined in the study will result in the department spending millions of Rands in repairing and replacing what has been lost in schools. This lose will be due to theft and vandalism, high rate of teachers resigning because of stress-related problems. In addition, the economy will be characterised by a high rate of unemployment due to learners who drop out of school, overcrowded classes will continue to affect teaching and learning. Ultimately, there will be a continuous decrease in the NSC results of entire District of Amathole West. For further or future research studies, the researcher suggested that studies on the department’s policies on school violence are needed. These studies can focus on what the policy entails, and the implementation and effectiveness of such policies. In addition to these studies, a study that closely looks at parents as contributing factors to school violence is needed. , Thesis (MEd) -- Faculty of Education, 2022
- Full Text:
- Date Issued: 2022-11
- Authors: Poswa, S
- Date: 2022-11
- Subjects: School violence , Violence in children
- Language: English
- Type: Master's theses , text
- Identifier: http://hdl.handle.net/10353/26075 , vital:64788
- Description: The purpose of this study was to look at the effects of school violence on teaching and learning in three selected schools in the Amathole West education district. The objectives of the study guided the entire project and the theory of socio-ecology by Benbenishty and Astor (2005) underpinned and guided the study. Different forms of school violence, causes of school violence, its effects and measures to curb its spread were discussed. The study made use of the pragmatic paradigm which allowed the researcher to make use of both quantitative and qualitative approaches. The data collection instruments consisted of questionnaires and semi-structured interviews. Three high schools were selected purposively with a sample of 75 respondents. Three principals, five teachers (school management team) and 7 learners (school representative council) participated in the semi-structured interview and 60 learners were randomly selected to take part in the questionnaire completion. Data analysis included coding of questionnaires and running the codes on the SPSS package. There after, results were presented using tables and descriptive analysis. Percentages and numeric figures of different forms of school violence were analysed. Data collected from the semi-interviews was coded according to different themes and categories and presented in form of discussions. Thematic analysis was applied in this regard. Therefore, the themes that emerged from the findings of the study were linked to the reviewed literature and research questions. The researcher cross-checked with the participants whether their responses were captured correctly before drawing any conclusions. The major findings of the study revealed that school violence is very much prevalent in the three selected schools. However, the severity of the problem varies from one school to another. Learners indicated that they experience bullying, verbal abuse, corporal punishment, cyberbullying, theft of property, gang-related violence and sexual harassment. Bullying, verbal abuse, cyberbullying and corporal punishment were rated as the most common and frequent forms of violence. In addition to the findings above, some causes of school violence noted by the respondents included peer pressure, poor parental care, exposure to violence at home or communities, unguided exposure to media, easy access to drugs, alcohol and weapons and teacher absenteeism. Effects of school violence included loss of concentration in class, poor academic results, class disruptions and depression or suicidal thoughts. All participants were allowed to say what they think should be done to curb the spread of school violence. Reasons given included that of ensuring that other departments are included in the committees that deal directly with school violence cases, strict and consistent enforcement of the school code of conduct and the department’s policies on school violence. Most learners and teachers indicated that parents are not present in the lives of their children. Thus, parental involvement was strongly recommended by most participants. The researcher had some recommendations to make. It was recommended that communities should take an active part in fighting drug and alcohol abuse by learners in communities. This could be done by reporting people who sell these drugs and alcohol to persons who are under the age of 18. It was further recommended that the department of education should understand and take advantage of technology. School violence apps can be developed specifically for teenagers. These will ensure that knowledge about school violence is spread to most learners. Lastly, workshops should be conducted with both teachers and learners to equip them more about school violence. The conclusion of the study highlighted some important points to be taken into consideration. One of the points was that a failure to implement the strategies outlined in the study will result in the department spending millions of Rands in repairing and replacing what has been lost in schools. This lose will be due to theft and vandalism, high rate of teachers resigning because of stress-related problems. In addition, the economy will be characterised by a high rate of unemployment due to learners who drop out of school, overcrowded classes will continue to affect teaching and learning. Ultimately, there will be a continuous decrease in the NSC results of entire District of Amathole West. For further or future research studies, the researcher suggested that studies on the department’s policies on school violence are needed. These studies can focus on what the policy entails, and the implementation and effectiveness of such policies. In addition to these studies, a study that closely looks at parents as contributing factors to school violence is needed. , Thesis (MEd) -- Faculty of Education, 2022
- Full Text:
- Date Issued: 2022-11
The experiences of students with a physical disability at institutions of higher learning: a systematic review
- Makubalo, Siphokazi https://orcid.org/0000-0001-8596-7357
- Authors: Makubalo, Siphokazi https://orcid.org/0000-0001-8596-7357
- Date: 2022-11
- Subjects: People with disabilities -- Education (Higher) , People with disabilities in higher education , South Africa
- Language: English
- Type: Master's theses , text
- Identifier: http://hdl.handle.net/10353/27242 , vital:66484
- Description: This study aims to explore the experiences of physically disabled students at institutions of higher. The research methodology followed in conducting this study is a systematic review. Thus, published literature from the period of 2000-2020 on the experiences of physically disabled students in South African institutions of higher learning was reviewed. Five articles which met the inclusion criteria were assessed for eligibility using the CASP criteria. The social model of disability is applied as a theoretical framework in exploring these experiences. A synthesis of results and findings of South African studies relating to physically disabled students in institutions of higher learning is provided. Themes that emerged in the study indicate that physically disabled students experience social exclusion related to architectural access barriers, attitudinal barriers, lack of support and discrimination in South African Institutions of higher learning, which impacts their overall physical access in these institutions. The study found that addressing these challenges and access barriers would result in positive experiences for physically disabled students on campuses. The study concluded that to improve the experiences of physically disabled students, institutional effort is necessary that is extensive both in scope and systematic in nature. For future studies, it is recommended that a qualitative study explore not just disabled students‟ lived experiences but also be inclusive of academic and support staffs‟ perceptions. , Thesis (MSoc) -- Faculty of Social Sciences and Humanities, 2022
- Full Text:
- Date Issued: 2022-11
- Authors: Makubalo, Siphokazi https://orcid.org/0000-0001-8596-7357
- Date: 2022-11
- Subjects: People with disabilities -- Education (Higher) , People with disabilities in higher education , South Africa
- Language: English
- Type: Master's theses , text
- Identifier: http://hdl.handle.net/10353/27242 , vital:66484
- Description: This study aims to explore the experiences of physically disabled students at institutions of higher. The research methodology followed in conducting this study is a systematic review. Thus, published literature from the period of 2000-2020 on the experiences of physically disabled students in South African institutions of higher learning was reviewed. Five articles which met the inclusion criteria were assessed for eligibility using the CASP criteria. The social model of disability is applied as a theoretical framework in exploring these experiences. A synthesis of results and findings of South African studies relating to physically disabled students in institutions of higher learning is provided. Themes that emerged in the study indicate that physically disabled students experience social exclusion related to architectural access barriers, attitudinal barriers, lack of support and discrimination in South African Institutions of higher learning, which impacts their overall physical access in these institutions. The study found that addressing these challenges and access barriers would result in positive experiences for physically disabled students on campuses. The study concluded that to improve the experiences of physically disabled students, institutional effort is necessary that is extensive both in scope and systematic in nature. For future studies, it is recommended that a qualitative study explore not just disabled students‟ lived experiences but also be inclusive of academic and support staffs‟ perceptions. , Thesis (MSoc) -- Faculty of Social Sciences and Humanities, 2022
- Full Text:
- Date Issued: 2022-11
The design of a training centre for female Artisans in Motherwell, Gqeberha
- Authors: Kelsley, Chanel
- Date: 2021-11
- Subjects: Continuing education centers -- South Africa -- Port Elizabeth -- Designs and plans , Artisans -- Training of -- South Africa -- Port Elizabeth
- Language: English
- Type: Master's theses , text
- Identifier: http://hdl.handle.net/10948/58694 , vital:60003
- Description: The treatise deals with the design of an Artisans’ Training Centre for the women in thetownship of Motherwell, Port Elizabeth. The aim of the building is to provide a safe civic space for teaching the women of Motherwell the necessary artisan skills. By teaching the women these skills, pov - erty amongst women in current township conditions can be improved by providing them with a sustainable income. Furthermore, the aim is to promote a sense of unity amongst these women as they enter a male-dominated field. South Africa’s skills shortages are widely regarded as key factors preventing the achievement of targeted economic growth rates. The decline and massive shortage in artisan trades has been attributed to the sharp decline of the apprenticeship system and the failure of substitute inter - ventions. By getting more women involved in the artisans’ field it will assist in gender equality and the current skills shortages. The treatise explores the issues and methods involved in the design of an Artisans Training Cen - tre for women. The research starts with an investigation into the nature of public space and civic facilities within a South African township context and how to create a more suitable environment for women, before exploring the development of an educational typology, while exploring the is - sues and processes involved in the design of a civic building in a township environment. The archi - tectural design is seen as an appropriate response to the identified architectural issues. Relevant precedents are chosen based on the nature of these issues and the context in which they are situated. The buildings are analysed in terms of these key issues and will thereby provide an understanding of the architectural identity of the civic and educational building archetype. An understanding of the architectural typology and an analysis of relevant precedents lead to the identification of a set of site selection criteria which can be used to select an appropriate site within the Motherwell precinct. The nature of Motherwell and the identified issues related to the selected site are then investigated in order to develop a set of informants and constraints. Specific elements which are unique to the selected site are then investigated individually while using relevant precedents to establish an appropriate response to the issues involved. An appropriate architectural response is subsequently developed through an understanding of the physical and spatial qualities of the building typology while also responding to the specific issues related to the site. , Thesis (MArch) -- Faculty of Engineering, the Built Environment and Technology, School of Architecture, 2021
- Full Text:
- Date Issued: 2021-11
- Authors: Kelsley, Chanel
- Date: 2021-11
- Subjects: Continuing education centers -- South Africa -- Port Elizabeth -- Designs and plans , Artisans -- Training of -- South Africa -- Port Elizabeth
- Language: English
- Type: Master's theses , text
- Identifier: http://hdl.handle.net/10948/58694 , vital:60003
- Description: The treatise deals with the design of an Artisans’ Training Centre for the women in thetownship of Motherwell, Port Elizabeth. The aim of the building is to provide a safe civic space for teaching the women of Motherwell the necessary artisan skills. By teaching the women these skills, pov - erty amongst women in current township conditions can be improved by providing them with a sustainable income. Furthermore, the aim is to promote a sense of unity amongst these women as they enter a male-dominated field. South Africa’s skills shortages are widely regarded as key factors preventing the achievement of targeted economic growth rates. The decline and massive shortage in artisan trades has been attributed to the sharp decline of the apprenticeship system and the failure of substitute inter - ventions. By getting more women involved in the artisans’ field it will assist in gender equality and the current skills shortages. The treatise explores the issues and methods involved in the design of an Artisans Training Cen - tre for women. The research starts with an investigation into the nature of public space and civic facilities within a South African township context and how to create a more suitable environment for women, before exploring the development of an educational typology, while exploring the is - sues and processes involved in the design of a civic building in a township environment. The archi - tectural design is seen as an appropriate response to the identified architectural issues. Relevant precedents are chosen based on the nature of these issues and the context in which they are situated. The buildings are analysed in terms of these key issues and will thereby provide an understanding of the architectural identity of the civic and educational building archetype. An understanding of the architectural typology and an analysis of relevant precedents lead to the identification of a set of site selection criteria which can be used to select an appropriate site within the Motherwell precinct. The nature of Motherwell and the identified issues related to the selected site are then investigated in order to develop a set of informants and constraints. Specific elements which are unique to the selected site are then investigated individually while using relevant precedents to establish an appropriate response to the issues involved. An appropriate architectural response is subsequently developed through an understanding of the physical and spatial qualities of the building typology while also responding to the specific issues related to the site. , Thesis (MArch) -- Faculty of Engineering, the Built Environment and Technology, School of Architecture, 2021
- Full Text:
- Date Issued: 2021-11