A method for extracting alarm thresholds from monitoring data based on normality model

Abstract

The invention discloses a method for extracting an alarm threshold value of monitoring data based on a normal model. , fit the remaining monitoring data after separating the periodic component set as a long-term trend; use the corresponding cumulative probability fitting function to fit the first monitoring data, the significant periodic set, and perform K-S fitting on the cumulative probability distribution model Goodness of fit test, based on the test results to obtain the optimal cumulative probability distribution model; based on the obtained monitoring data, the periodic component set, and the minimum probability value of the long-term trend, the extrapolated probability interval of the monitoring data is constructed, and the small probability of the extrapolated interval is selected. value to construct an alarm threshold based on a small probability value. The method can accurately warn the long-term change trend and change cycle of the monitoring data of cultural relics.

Description

technical field [0001] The invention belongs to the field of data mining, in particular to a method for extracting alarm thresholds of monitoring data based on a normal model. Background technique [0002] Immovable cultural relics refer to cultural relics with important scientific value and actual existence in history, culture, architecture and art, including traditional historical sites such as ancient buildings and settlements, covering politics, military, production activities, sacrifice, art, society, construction, education, etc. [0003] The protection of immovable cultural relics refers to the process of adopting various means to prevent cultural relics with important historical, artistic and scientific value from being damaged. In recent years, my country has carried out a large number of targeted research and implementation work. Some achievements have been made in this regard, and the weathering of a number of key cultural relics has been significantly curbed and ...

AI Technical Summary

Problems solved by technology

There are two problems with this method: one is that the method is based on the assumption that the sensor monitoring data obey the normal distribution under normal conditions, but in fact most of the sensor monitoring data do not conform to the normal distribution; The sensor monitoring data itself gives an alarm, and there is no in-depth analysis and alarm on its long-term trend and change cycle

Images