The paper suggests a method for analyzing technology trends. The process, which investigates development of technologies over time, identifies main technologies displaying the fastest growth compared to greater influence of new inventions. The method analyzes term frequency and change over time of technological terms in academic articles and patents to identify the prior technologies that lead to a new technology and to detect technologies that have the biggest impact. The analysis was performed on 4,354,054 patents from the US Patent Office dating from 1975 until today. In addition, academic articles were analyzed as a trend forecasting dataset to identify patents trends 4-5 years in advance and technology trends up to 9 years in advance. The forecasting method was extensively validated using a large repository of real-world technology terms, and the results were verified against Gartner technology predictions, Web searches, news articles, and book publications. The method shows higher accuracy than existing forecasting methods do. Some correlation is displayed between technology trends and future US stock market performance.