The field of data mining has seen an explosion of interest from both academia and industry and this interest continues to grow at a rapid rate. In this context, we use the term data mining to refer to all aspects of an automated or semi-automated process for extracting previously unknown and potentially useful knowledge and patterns from large databases. This process typically involves several steps ranging from data collection and integration to interpretation of the output, but we note in particular that the success of the process relies heavily on the availability of computer algorithms that can be effectively applied to large data sets to extract useful information.
The operations research community has recently made significant contributions in this area and in particular to the design and analysis of data mining algorithms. For example, mathematical programming formulations of support vector machines have been used for feature selection and data clustering. Metaheuristics and evolutionary methods have also been introduced to solve these and other data mining problems, and the opportunities for using both exact and heuristic optimization algorithms for data mining are extensive. This includes but is not limited to data mining problems such as feature and instance selection, classification, association rule discovery and data clustering, and OR methodologies such as mathematical programming, evolutionary methods, and metaheuristics.
However, the intersection of OR and data mining is not limited to algorithm design and data mining can play an important role in many OR applications. Applications in areas such as e-commerce, transportation and logistics, supply chain management, planning and scheduling, and inventory control often generate vast amounts of data and data mining can be used to extract structural information and insights from these datasets. An example of this that has been studied in the past is the use of data mining to learn when a particular dispatching rule is appropriate in production scheduling environments.
Finally, while the data mining process is usually effective for generating insights and patterns from large data sets, it is a model free approach and the insights are typically unstructured and require substantial interpretation. Thus, optimization methods can potentially be applied to the output of the data mining process to optimize the desired objective while accounting for relevant business constraints. Again, all of the methodologies or application areas mentioned above are relevant in this context as well.
This focused issue aims to broadly explore the synergy between the fields of operations research and data mining. We are therefore interested in receiving papers focusing on either applications or methodology and dealing with any aspect of the intersection between these two fields, including but not limited to the three perspectives described above.
The deadline for submission is December 31, 2003 but early submissions are encouraged. Prospective authors should send a postscript (.ps) or pdf formatted file containing their paper as an email attachment to the guest editor at firstname.lastname@example.org. Any questions regarding this focused issue should also be directed to the guest editor. All submitted papers will be peer reviewed according to the usual standards of a leading international journal.
General information and author guidelines can be found at the website of Computers & Operations Research.