- Title
- Classification and clustering based methods for outlier detection of solar resource data
- Creator
- Abrahams, Waldo
- Subject
- Port Elizabeth (South Africa)
- Subject
- Eastern Cape (South Africa)
- Subject
- South Africa
- Date Issued
- 2022-04
- Date
- 2022-04
- Type
- Master's theses
- Type
- text
- Identifier
- http://hdl.handle.net/10948/55423
- Identifier
- vital:51996
- Description
- Almost 90% of the primary global energy demand is serviced from the burning of fossil fuels (Abas, Kalair & Khan, 2015). Owing to the detrimental environmental impact of this, a global energy transition to the use of renewable energy, including solar energy, is needed (Gielen et al., 2019). An important aspect that inhibits the growth of solar energy is accurate solar resource data. Such data is needed because knowledge of the future reliability and quality of energy production is required to analyse a system’s performance and determine financial implications (Sengupta et al., 2017). Existing methods used to detect outliers in solar resource data do not efficiently identify outliers and an accurate and robust approach is required (Eastwood, 2019). Using simulated and real-world data, this study investigates the use of several classification methods, along with a two-stage clustering-classification approach to accurately identify outliers in solar resource data. The Treebag method proves to be an adequate outlier detection method for solar resource data.
- Description
- Thesis (MSc) -- Faculty of Science, School of Computer Science, Mathematics, Physics and Statistics, 2022
- Format
- computer
- Format
- online resource
- Format
- application/pdf
- Format
- 1 online resource (124 pages)
- Format
- Publisher
- Nelson Mandela University
- Publisher
- Faculty of Science
- Language
- English
- Rights
- Nelson Mandela University
- Rights
- All Rights Reserved
- Rights
- Open Access
- Hits: 3437
- Visitors: 3461
- Downloads: 88