Share
  • This paper introduces a method using a Bayesian network to aggregate trends detected in time-series data with events identified by natural language processing to improve the accuracy and robustness of kick and lost-circulation detection.