CHAPTER 7
SUMMARY AND FUTURE RESEARCH
We believe that the work presented in this document has demonstrated the potential of the research program we call OCCURRENCE-BASED processing. Although quite simple in concept, it appears to be producing interesting results. Therefore, we believe the occurrence-based methodology warrants additional study.
Below, we will first summarize the key results from the preceding chapters. In the process, we will re-emphasize the technique's strengths. It is important to emphasize that this work is very preliminary in nature. Very little effort has been placed on technique optimization. Our work differs from existing statistical approaches to word categorization [see SCHUTZE93, FINCH92, BRILL92] because we have ignored frequency counts. The goal of this work was to determine whether the technique could produce interesting results. We believe that our work shows that it does.
The preceding chapters also identified a number of the technique's weaknesses. In many cases, what appear to be weaknesses are simply areas that require additional research. In the second half of this chapter, we will expand on a number of these areas. We will mention a number of possible research directions that may address these problems. We believe that the occurrence-based methodology shows great potential, and these research projects should be pursued.