Offshore facility's fire risk estimation method using machine learning

摘要

The method for predicting fire risk of offshore facilities using machine learning according to an embodiment of the present invention includes receiving design data of existing offshore facilities, which is design data of previously designed offshore facilities, and data characteristics from the inputted design data of existing offshore facilities. Extracting, generating a fire risk of the previously designed marine facility as a data label, generating the data characteristic and the data label as a data set, the generated data set as a training data set and verification data Separating into three, performing a machine learning analysis through the training sets to obtain a functional equation functionalizing the correlation between the data characteristic and the data label, and determining the fire risk of the offshore facility through the obtained functional equation. Generating a prediction model for prediction, verifying the generated prediction model using the verification data set, and inputting new offshore facility design data into the generated prediction model to determine the fire risk of the new offshore facility And predicting.