| 摘要 |
<p>This invention relates in general to database management systems performed by computers, and in particular, to database management systems that support on-line analytical processing (OLAP). The present invention teaches a means and a computerized method for automatically determining one or more variable dimensions for a multi-dimensional database table, whereby the multi-dimensional database table also comprises at least one pre-defined measure dimension for storing values of one or more measure types. The current invention suggests to treat for the purpose of determination of the variable dimensions a multitude of input records as an implicit, yet unknown functional relationship between the measure types (as a dependent variables) and the variable types (as independent variables). Based on this insight it is then suggested to calculate a regression function as a prediction model for the (dependent) measure types based on the (independent) variable types utilizing the multitude of input records. The most significant variable types contributing to this prediction model are then selected as variable dimensions for the multidimensional database table, i.e. as the OLAP cube dimensions.</p> |
