CAMeL: A Self-Adaptive Framework for Enriching Context-Aware Middlewares with Machine Learning Capabilities

Figure 5

Precision (a) and recall (b) of vehicle classification () obtained with or without automata-driven reconfiguration. The average throughput (c) of classification components decreasing linearly with an increasing number of components. The heap memory used by our framework (d) increasing linearly with the number of messages processed. Both curves have been computed for both desktop and mobile architectures.

(a)

(b)

(c)

(d)