Hierarchy-Based Update Propagation in Decision Support Systems

Sales forecasting systems are used by enterprise managers and executives to better understand the market trends and prepare appropriate business plans. These decision support systems usually use a data warehouse to store data and OLAP tools to visualize query results. A specific feature of sales forecasting systems regarding future predictions modification is backward propagation of updates, which is the computation of the impact of modifications on summaries over base data. In Data warehouses, some methods propagate updates in hierarchies when data sources are subject to modifications. However, very few works have been devoted so far, regarding update propagation from summaries to data sources. This paper proposes an algorithm called PAM (Propagation of Aggregate Modification), to efficiently propagate modifications on summaries over base data. Experiments on an operational application (Anticipeo) have been conducted.

[1]  Haitang Feng Performance Problems of Forecasting Systems , 2011, ADBIS.

[2]  Inderpal Singh Mumick,et al.  Selection of Views to Materialize in a Data Warehouse , 2005, IEEE Trans. Knowl. Data Eng..

[3]  Alberto O. Mendelzon,et al.  Maintaining data cubes under dimension updates , 1999, Proceedings 15th International Conference on Data Engineering (Cat. No.99CB36337).

[4]  Kenneth A. Ross,et al.  Complex Aggregation at Multiple Granularities , 1998, EDBT.

[5]  Yvan Bédard,et al.  Handling evolutions in multidimensional structures , 2003, Proceedings 19th International Conference on Data Engineering (Cat. No.03CH37405).

[6]  Hicham G. Elmongui,et al.  Lazy Maintenance of Materialized Views , 2007, VLDB.

[7]  Dan Suciu,et al.  Adding Structure to Unstructured Data , 1997, ICDT.

[8]  Dimitri Theodoratos Exploiting hierarchical clustering in evaluating multidimensional aggregation queries , 2003, DOLAP '03.

[9]  Isidro Ramos,et al.  Advances in Database Technology — EDBT'98 , 1998, Lecture Notes in Computer Science.

[11]  Chaman L. Jain Editorial: Which Forecasting Model Should We Use? , 2000 .

[12]  Elke A. Rundensteiner,et al.  GPIVOT: efficient incremental maintenance of complex ROLAP views , 2005, 21st International Conference on Data Engineering (ICDE'05).

[13]  Elke A. Rundensteiner,et al.  Multiversion-based view maintenance over distributed data sources , 2004, TODS.