Investigating sharing of information between nodes for improved prediction

Project on sharing data for distributed machine learning energy prediction algorithm (EH-WSN).

Built on top of existing energy prediction algorithm (EWMA, QL-SEP).
Tried several methods of sharing data between sensor nodes.
Tried various types of data to help improving prediction.
Varied the time of sharing, how frequent the sharing occurs in a day.

Final result : Not a good method.

• Less than 5% improvement in prediction accuracy.
• High energy consumption for sharing data as compared to conduct extra sampling.