Stochastic technology life cycle analysis using multiple patent indicators

Abstract

Technology life cycle analysis plays a crucial role in setting up investment-related strategies. The dominant approach to technology life cycle analysis utilizes curve fitting techniques to observe technological performance over time. However, doubts have been expressed about the accuracy and reliability of this method, due to its use of single indicators and the necessity of making assumptions about pre-determined growth curves. As a remedy, we propose a stochastic technology life cycle analysis that uses multiple patent indicators to examine a technology's progression through its life cycle. We define and extract seven time-series patent indicators from the United States Patent and Trademark Office database, and employ a hidden Markov model-which is an unsupervised machine learning technique based on a doubly stochastic process-to estimate the probability of a technology being at a certain stage of its life cycle. Based on this model, this paper also investigates patterns of technology life cycles, future prospects of a technology's progression, and characteristics of patent indicators between technology life cycle stages. The systematic process and quantitative outcomes the proposed approach offers can facilitate responsive and objective technology life cycle analysis. A case of molecular amplification diagnosis technology is presented.