Abstract | ||
---|---|---|
Predicting failure or success of an event or value is a problem that has recently been addressed using data mining techniques. By using the information we have from the past and the information of the present, we can increase the chance to take the best decision on a future event. In this paper, we evaluate some popular classification algorithms to model a water quality detection system. The experiment is carried out using data gathered from Thuringer Fernwasserversorgung water company. We briefly introduce baseline steps we followed in order to achieve a descent model for this binary classification problem. We describe the algorithms we have used, and the purpose of using each algorithm, and in the end we come up with a final best model. Representative models are compared using the F1 score, as a performance measurement. Finding the best model allows for early recognition of undesirable changes in the drinking water quality and enables the water supply companies to counteract in time. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1007/978-3-319-76081-0_15 | Studies in Computational Intelligence |
Keywords | Field | DocType |
Classification,Watter quality,Performance metrics | F1 score,Data mining,Binary classification,Computer science,Performance measurement,Artificial intelligence,Statistical classification,Water quality,Machine learning,Water supply | Conference |
Volume | ISSN | Citations |
769 | 1860-949X | 2 |
PageRank | References | Authors |
0.42 | 3 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Fitore Muharemi | 1 | 3 | 1.13 |
Doina Logofatu | 2 | 17 | 16.74 |
Christina Andersson | 3 | 2 | 0.76 |
Florin Leon | 4 | 71 | 15.03 |