Implicit Customer Relations Mining with the Event-Indexing Model

Exploration of explicit and, especially, implicit information of customers is a vital task for industrial enterprises and organizations. Text and data mining are two indispensable means to accomplish above task. This paper presents a hybrid mining system, EIREx, which processes unstructured (text documents) and structured (data tables from databases) information as a whole source. The construction of EIREx is inspired by a cognitive situation model, event-indexing model. EIREx monitors customer information in multidimensional vectors to simulate cognitive understanding processes of human beings. To ensure accuracy, EIREx implements syntactic parsing,lexical reference and semantic analysis in information extraction procedures from mining sources. In the procedures of information integration and cross referencing,EIREx delivers a structural description to reveal implicit connections among existing customers, which are valuable for further analysis of customer orders and purchase motivations or intentions, and dealing with customers´ complaints and suggestions.