Co-occurrence Analysis In Applied Research, Knowledge Mining In The Text

With the dramatic increase of data presented in text format and users' need of mining knowledge within it, the study of text-knowledge is popping up. Due to the differences between text and ordinary records from database, it's difficult to use computer to deal with it. in order to obtain a comprehensive and embedded understanding of contents, We have to find a method which is different from traditional data mining method and pertinent to the characteristic of text. Therefore, it's of vital importance to discuss the theory of Co-occurrence analysis and its concrete application in text-knowledge mining.First of all, this paper discussed the definition, ordinary procedure, basic method and major issue of text-knowledge mining. And then the author discussed the theory of Co-occurrence analysis, including its definition, category, methodology and the common annalistic procedure. On the basis of all above, the function of utilizing Co-occurrence analysis in text-knowledge mining was proposed: it provided the courses of text-knowledge mining with semantic support, it discovered interesting knowledge patterns from the association of terms, and it serves as an effective method to discover text-knowledge. Then, based upon the third function, the author brought forward the way it was used in text-knowledge mining research from three perspectives—based on spatial distribution, based on temporal distribution and based on the integrated mapping of outer and inner information in text. In order to illustrate, some representative cases were introduced as well; the applicable range of its utilization and a common analytic procedure were put forward, and major problems within the procedure were discussed in depth; In the next place, by using the above research results as a kind of methodology, and by using the data stored in the database of magazine literature of China under the category of 'aero engine', a exploring application was carried through. Its findings were that, the hot topics of aero engine included flaming system, control, depressor, watering tube, and turbine. The flaming system remained as the core issue of the field, and other topics all had a trend of getting closer to the core through widening the scope and boosted up depth of research. In the near future, the structure of aero engine won't change much; finally, the new trend of using Co-occurrence analysis method in text-knowledge mining was summarized.