2015-11-06 Towards Declarative Scripting Combining CP and Analytics
6 novembre 2015, salle 3/124 du Bâtiment 5 du LIRMM
par Hassan Aït-Kaci (LIRIS - Université Claude Bernard Lyon 1)
Résumé: We take a rapid look at the current trends of evolution of application software tools for combining the formal capacities of (mostly statistical) Analytics along with Constraint Programming (CP). We note that application orchestration is what makes this task difficult. As a result, a new area of software development is rapidly becoming of great importance in relation with CP and Analytics: viz., scripting application orchestration. This has become a clear need in particular for Data Analytics that are typically multiplexed applications interacting between data and specific processing tools acting thereon such as statistical analysis and simulation, optimization, graphing, tabling and DB interface, HTML/LaTeX/Doc report generation, etc., ... We overview some systems used for scripting application orchestration involving CP and Analytics, looking at pros and cons. Then, we discuss the following point: since scripting is crucial both to CP and Analytics, shouldn't we leverage CP to make Analytics programming as declarative as possible? If so, what are the challenges?
This was an invited presentation as part of the CP-2015 Panel Workshop on CP and Analytics, Aug. 31, 2015, Cork, Ireland.
Dernière mise à jour le 20/10/2015