A Forecasting Model for Technological Trend Using Unsupervised Learning

This paper proposes a forecasting model for technological trend using unsupervised learning, and uses association rule mining and self organizing map as un supervised learning methods for image and video technology.

Abstract

Many results of the developed technologies have applied for patents. Also, an issued patent has exclusive rights granted by a government. So, all companies in the world have competed with one another for their intellectual property rights using patent application. Technology forecasting is one of many approaches for improving the technological competitiveness. In this paper, we propose a forecasting model for technological trend using unsupervised learning. In this paper, we use association rule mining and self organizing map as unsupervised learning methods. To verify our improved performance, we make experiments using patent documents. Especially, we focus on image and video technology as the technology field.