Domain Knowledge in Predictive Maintenance for Water Pipe Failures
- Publisher:
- Springer
- Publication Type:
- Chapter
- Citation:
- Human and Machine Learning, 2018, pp. 437-457
- Issue Date:
- 2018-06-08
Closed Access
Filename | Description | Size | |||
---|---|---|---|---|---|
978-3-319-90403-0_21.pdf | Published version | 689.1 kB |
Copyright Clearance Process
- Recently Added
- In Progress
- Closed Access
This item is closed access and not available.
In this chapter, the water pipe failure prediction is used as an example to show the integration of machine learning and domain knowledge. It is crucial for the risk management strategy of water distribution systems to minimise the water pipe failure impacts. Prediction of water pipe conditions through statistical modelling is an important element for the task. When applying the models to practical problems, domain experts can provide invaluable suggestions that can be used as constraints or informative prior knowledge. Alternatively, the models can also help domain experts to explore more insights. The chapter uses major steps in the water pipe failure prediction, including data review, factor analysis, prediction evaluation and practical use, as examples to illustrate how the domain knowledge is integrated. Then the hierarchical non-parametric model is used as an example model
Please use this identifier to cite or link to this item: