A Text Representation Method Based on Harmonic Series

Text representation problem is one of the fundamental but most important issues in the areas of information processing and retrieval, and still a big challenge. From the perspective of cognitive psychology, this paper employs the ideas of the human cognitive process to the machine text understanding. And aiming at the structure of paper texts, we propose a cognition network model based on the Harmonic Series with the consideration of both key occurrence frequencies and occurrence locations. The proposed model is not only simple, but also effective and can be easily handled as RFID data in the Internet of Things.