Review of Outlier Detection Algorithms Based on Grain Storage Temperature Data

Abstract

Outliers may be objectively existing data generated by mechanisms different from most data, and have related laws and certain research value. Outliers have been widely used in recent years, including financial fraud, cyber fraud, and abnormal weather warnings. Traditional outlier detection algorithms such as statistical-based outlier detection, cluster-based outlier detection, deviation-based outlier detection, distance-based outlier detection, density-based outlier detection, etc., but there are many shortcomings in the traditional outlier detection algorithms. Outlier detection algorithms based on information entropy, outlier detection algorithms based on spectral clustering, etc. have gradually appeared. Because in practice, traditional outlier detection algorithms cannot perform good detection applications on high-dimensional data, spatial data, and time series data, so they are performed for high-dimensional outlier detection, space outlier detection, and time series outlier detection. Dedicated research. How to improve the effectiveness, scalability and space-time efficiency of the algorithm will be the future research direction.