Utilizing ARM Technique in Mining Textual Data

Text mining, as one major school in Knowledge Discovery in Data (KDD), mines hidden patterns, rules, regularities and trends from textual data / non-database-data (i.e., text files, web documents, etc.). It is quite different from data mining (another well-known major school in KDD): the data structure of texts, dealt by text mining, is considered implicit, whereas traditional database-data, dealt by data mining, is relatively explicit / structured. With the appearance of various approaches in text preprocessing, “rich” texts can be converted to useful structured / semi-structured data. Hence, it is arguably doable to obtain knowledge from texts by utilizing the traditional data mining techniques, such as classification rule mining, clustering, Association Rule Mining (ARM), etc. In this paper, we summarize the wide usages of utilizing ARM technique in the field of text mining, Text Association Rule Mining (TARM). Our work is presented with the aim of supporting future work in text mining research.