We are happy to introduce a visual analytics approach based on an interactive visualization of time series data involving machine-learning approaches for anomaly identification in time series data of hydrocarbon reservoir production.
The methods leverage the prior probability of anomalies from a time window, do not require labeled training data with normal and abnormal conditions, and incorporate specialist knowledge in the exploration process. Anomalies can result for different reasons: gross errors, system availability, human intervention, or abrupt changes in the series.
Paper freely available at: https://lnkd.in/eSWjgpU !
Our methods detect approximately 95% of the human intervention anomalies and about 82%–89% detection rate for other anomalies identified during data exploration.
In collaboration with Aurea Soriano-Vargas Pedro Ribeiro Mendes Júnior Alexandre Mello Ferreira Rafael de Oliveira Werneck Renato Moura Raphael Felipe Prates Manuel Castro Maiara Moreira Gonçalves Manzur Hossain Marcelo Ferreira Zampieri Alessandra Davolio Denis Schiozer (all at Unicamp) and Bernd Hamann (at Univ. of California, Davis).
This work is a partnership with Shell !