The abductive inference of the drilling process based on the multimedia data about the drilling equipment

Abstract

The main objective of the proposed research is to build up the methodology for combining data mining technologies with the framework of logic programming with constraints that allow the integration of domain knowledge’s about the drilling process into the overall structure of the data and knowledge’s mining for the oil and gas domain with the possibilities of direct application with the selected classifier, particularly, in the form of decision tree that make use of crispy and posibilistic-based level of framework application that basically implements the abductive methods of inference construction. The used research methods can be reduced to the pure concept of the abductive framework implementation which at the very input receives the data sets in the form of the graphic markers that correspond to the specific levels or meta-levels for the domain knowledges in the form of raster images, which generally allows to build the computable solutions that perform effective verification of the initial hypotheses that were observed. The main result is that the executable application classifier in the form of a tree structure of the constraints network allows the use of direct abductive computing to the new observations with the purpose of achieving the appropriate threshold number of rules presented in the form of «relevancy condition –satisfaction condition" that, in turn, greatly expands the scope of abductive computing as a whole, particularly the level of multiple-valued logic. As the main result there should be highlighted the submitted application methodology of logic programming with constraints to the problems of extracting pure relational media or data sets that require the use of primary treatment by creating of the sequences of significant graphical markers based on descriptive tools of the first order, which generally allows the building of applications for the restricted search spaces based on marked and probabilistic values, which ultimately allows the extension of the original framework for the level of abductive probabilistic framework that will be based on the semantically founded extensions of first-order logic