A long short-term memory network (LSTM)-based astronomical big data optically variable curve abnormity detection method

Abstract

The invention discloses a long short-term memory network (LSTM)-based astronomical big data optically variable curve abnormity detection method. The method aims to solve the problem that the abnormityof an optically variable curve in astronomical data is difficult to detect due to the instantaneity of the abnormity. According to the method, a large amount of optical variable curve data obtained through observation of a GWAC serve as a training sample and a test sample, meanwhile, a neural network formed by linking LSTM units is constructed, and a network model is trained continuously, so thatan output value obtained in a special mode of the network is obtained. According to the method, the weight of each unit is continuously updated through iterative training, so that the overall situation is optimal. After the training is finished, the next prediction value of the sample sequence can be obtained only by inputting the test sample into the network.

Description

technical field [0001] The invention belongs to the technical field of astronomical big data processing, and relates to a long-short-term memory network-based method for anomaly prediction of light curves of astronomical big data. Background technique [0002] Astronomical research often deals with a large number of large datasets generated and integrated through surveys. For example, the recent China-France joint astronomical satellite project SVOM. The main purpose of the project is to detect the phenomenon of gamma-ray bursts in the universe. The Ground Wide Angle Camera Array (GWAC) in the SOVM project is used to measure light data in the universe and generate a dataset of astronomical light curves. This system includes 36 wide-angle cameras, each camera has a diameter of 18 cm, a focal length of 22 cm, and a 4k*4k CCD detector. CCD detectors are very sensitive in the wavelength range from 400 to 800 nm. The camera array is capable of acquiring images in a total of o...

AI Technical Summary

Problems solved by technology

In view of the fact that there are fewer research methods on light curves than in other fields, and it is difficult to deal with the huge amount of data, it is necessary to propose a model based on the LSTM neural network model and rely on a large amount of data obtained by astronomical telescopes as the counterpart of the data set. A Method for Predicting Abnormality of Light Curve

Images