An adaptive approach with active learning in software fault prediction
Proceedings of the 8th International Conference on Predictive Models in Software Engineering
| Hi-index | 0.00 |
Combining adaptive heartbeat mechanism with fuzzy grey prediction algorithm, a novel implementation of failure detector is presented. The main parts of the implementation are adaptive grey prediction layer and adaptive fuzzy rule-based classification layer. The former layer employs a GM(1,1) unified-dimensional new message model, only needs a small volume of sample data, to predict heartbeat arrival time dynamically. Then, the predict value and the message loss rate in specific period are act as input variations for the latter layer to decide failure/non-failure. Furthermore, algorithms of how to predict arrival time and how to construct adaptive fuzzy rule-based classification system are presented. Experimental results validate the availability of our failure detector in detail.