Algorithm 1. Data fusion process.

1. When an event occurs, sensor nodes are awakened to perceive and collect event information;

2. for $T\ge 0$ do

The sensor nodes will conduct simple integration of the data set monitored in the t period, remove redundant and identical data;

while $D_{T_i}\ne D_{T_{i-1}}$do

The integrated data $D_{T_i}$ will be transmitted to its cluster- head sink node;

end while

3. end for.

4. The sink node will integrate and standardize the data collected after the T period;

5. The sink node obtains the weight of each attribute value according to the data fusion algorithm, and synthesizes the final value of each attribute value according to the attribute value;

6. The sink node merges the final value of each attribute value into the final prediction result $R_{final}$;

7. If $R_{final}$ are changed, do

The sink node will periodically transmit the calculated final prediction results to the base station.

8. Else

$R_{final}$ will not be transmitted;

9. end if

10. The base station will comprehensively evaluate the collected results to determine the possibility of the event happening $P_{final}$ in the monitoring area;

11. If $P_{final}$ are changed, do $P_{final}$ is transmitted to the terminal;

12. Else

13.   $P_{final}$ will not be transmitted.

14. end if