One way to cluster documents by meaning using semantic-ball

Semantic technology dealing with meaning is difficult to process meaning as intended by people due to it is flexibility, instability, and even uncertainty. Moreover, there is no explicit definition of how to sense meaning and expression by computer. To overcome these barriers, this paper suggests Semantic-ball, which is a co-occurrence of different word-set, and based on assumptions that Semantic-ball has emergence self-organization attribution, group relationship not pair relationship, and specificity. Therefore this paper focuses on finding Semantic-ball to solve semantic problems. Approach to find Semantic-ball is based on conventional statistical ways, the FP Growth algorithm(Frequency Pattern Growth algorithm)[1], and the bottom-up method. While the previous researches have been based on the definition which used to pin down meaning to solve semantic problems, this paper does not bring any definition. Moreover, Semantic-ball is made only using computing ability automatically. Thus, Semantic-ball helps to sense word and cluster documents by meaning.