Can Automatic Personal Categorization deal with User Inconsistency?

Document categorization is a daily task in every organization, but it is a very subjective process. While automatic document categorization has been widely studied, much challenging research still remains to support user subjective categorization. This study evaluates and compares the application of Self-Organizing Maps (SOM) and Learning Vector Quantization (LVQ) to automatic document classification according to a subjectively predefined set of clusters in a specific domain, and assesses the effect of user inconsistency on this process. Results show that despite the subjective and inconsistent nature of human categorization, automatic document clustering methods correlate well with subjective, personal clustering. Moreover, adapting a system to its users is limited by users' inconsistency, meaning that a perfect adaptation is an impractical goal.