Online Customer Review Analysis Technique based Relational Table with Constant Degree
Keun-Hyung Kim · 2011
In case of transforming online customer reviews into relational table, there exist many null values in the table because the lengths of each online customer reviews are different. The null values might bring both the delay of processing time and wastefulness of storage. In this paper, we proposed the novel techniques of decreasing the processing time and storage space for analyzing the online customer reviews. The basic idea of the technique is to fix the degree of relational table and reduce null values when transforming the online customer reviews to relational table. We implemented the prototype system for analyzing the reviews in order to evaluate how different the performances in calculating the frequencies of appearances of nouns are between types of the transformation tables. We confirmed that the transformation table with constant degree derived excellent performance. In particular, we recognized that the smaller the degree of the constant transformation table, the shorter the time in calculating the frequencies.