A combination of ROCCO system and Support Vector Machines in occupation coding
Kazuko Takahashi · 2004
We have used Rule-based OCcupation COding (ROCCO) system, which is a rule-based automatic method, to the occupational/industrial data of JGSS-2000/2001/2002. The performance of ROCCO system is more stable than that of coders and by using it, we can lighten work of coders and save much time. However, its categorization performance is not satisfiable. Therefore, we adopt Support Vector Machines (SVMs) , which show high performance in various fields, and compare it with the rule-based method. We also investigate effective combination methods of SVMs and ROCCO system, especially the method we use the results of SVMs if ROCCO can’t decide occupational code. We empirically show that SVMs outperform ROCCO system in the occupation coding and that the combination of the two methods yields an even better accuracy. We can improve the accuracy by contriving the combination of two methods.