Comparing compositional multivariate outliers with autoencoder networks in anomaly detection at Hamich exploration area, east of Iran
H Moeini, FM Torab - Journal of Geochemical Exploration, 2017 - Elsevier
Newly presented machine learning methods based on Deep Belief Networks like
autoencoders have opened a new window on anomaly identification in different fields of the
science. They reconstruct the normal probability distribution pattern of the input data using
stacks of Continuous Restricted Boltzmann Machines (CRBM) and thus determining the
outliers. Therefore using this machine on geochemical samples taken in regional
exploration scale, might be an acceptable way to delineate the multivariate anomalies and …
autoencoders have opened a new window on anomaly identification in different fields of the
science. They reconstruct the normal probability distribution pattern of the input data using
stacks of Continuous Restricted Boltzmann Machines (CRBM) and thus determining the
outliers. Therefore using this machine on geochemical samples taken in regional
exploration scale, might be an acceptable way to delineate the multivariate anomalies and …
以上显示的是最相近的搜索结果。 查看全部搜索结果