| CPC G06Q 50/265 (2013.01) [G06F 16/285 (2019.01); G06N 5/022 (2013.01)] | 11 Claims |

|
1. A report data convergence analysis method for a disaster response to a disaster, the report data convergence analysis method comprising:
receiving first input data including report input data of a reporter, reception input data of a receiver, and report document input data that is prepared by the receiver after a disaster situation ends;
storing the first input data in a first database;
expanding data required for learning by augmenting the first input data stored in the first database;
generating a plurality of first learning models through a learning process based on the first input data stored in the first database;
calculating disaster situation recognition accuracy indicating disaster situation recognition performance for each generated first learning model;
receiving second input data including the first input data and response data for each disaster situation;
storing the second input data in a second database;
further expanding the data required for learning by augmenting the second input data stored in the second database;
generating a plurality of second learning models through a learning process based on the second input data stored in the second database;
when the report input data of the reporter is received, performing convergence analysis on the plurality of generated first learning models to provide disaster situation information; and
when the report input data, the reception input data, and the disaster situation information are received, performing convergence analysis on the plurality of generated second learning models to generate further disaster response information;
wherein the generating of the plurality of first learning models through the learning process based on the first input data stored in the first database includes calculating a weight for convergence analysis of the first learning models based on the disaster situation recognition accuracy calculated for each generated first learning model; and
wherein the generating of the plurality of second learning models through the learning process based on the second input data stored in the second database includes calculating a weight for convergence analysis of the second learning models based on situation response suitability for each generated second learning model;
the report data convergence analysis method further comprising adjusting the weights of the first learning models and the weights of the second learning models according to the disaster situation recognition accuracy and the situation response suitability;
wherein the disaster situation information includes information required for a response to the disaster.
|