Recommendation systems − based on the functionalities clustering, classification, and prediction − automate information processing steps such as the classification of artifacts and assist the user in decision-making processes. Many of these systems are realized by symbolic methods such as association rule mining or by rule-based mechanisms in general. Consequently, they lack the capability of generating recommendations for new and hitherto unprocessed contents by means of generalization.
The hybrid and domain-independent framework developed in this dissertation called SymboConn is based on a connectionist learning algorithm and is, however, able to process symbolic knowledge. In addition, the framework provides a high generalization capability, flexibility, and robustness. We demonstrate its applicability by several case studies in the areas of navigation recommendation for web pages, design pattern discovery, change impact analysis for software development, and time series prediction for demand planning.
«
Recommendation systems − based on the functionalities clustering, classification, and prediction − automate information processing steps such as the classification of artifacts and assist the user in decision-making processes. Many of these systems are realized by symbolic methods such as association rule mining or by rule-based mechanisms in general. Consequently, they lack the capability of generating recommendations for new and hitherto unprocessed contents by means of generalization.
The...
»