IIDMCC: An Innovation Idea Discovery Model Using Online Customers' Complaint Messages

Abstract

Online customers’ complaints have attracted increasing attention to innovation developers. By applying text mining and classification-oriented data mining techniques, an innovation idea discovery model using online customers’ complaint messages (IIDMCC) was proposed and implemented in this article. Methods included text mining to derive bags of words, sparsity exclusion to produce a term matrix, and supervised classification data mining to reveal decision rules. The IIDMCC showed 90.63% prediction accuracy based on 14720 complaint messages collected from official forum and online communities of a case company in the mobile phone sector from Taiwan. Validation of data inputs, method, and outputs was conducted via case company specialists. The article concludes that analyses of online complaint messages may potentially contribute to the exploration and discovery of innovation ideas. The paper demonstrates the use of mining open textual data in general and complaint messages in particular in the domain of knowledge discovery in databases.