Patent data analysis with network analysis and time series methods

Both private and public institutions are very interested in prior knowledge of emerging technologies in order to make strategic investments and learn how to advance their work. Technology forecasting from patent documents to follow developments and follow trends is a recent and shining field of research. Until today, many studies have been conducted in different fields that make technology forecasts and share their results. However, studies in the field of technology forecasting for information security, which is one of the most important issues of today, are limited. Recently, information security has become a complex structure that concerns both software, hardware and people. As a result, new technologies emerging in the field of information security are increasing rapidly. In this study, among the patents obtained in the last decade, those concerning information security were prepared as a data set, and the patents in the data set were examined by network analysis and time series analysis methods. With the network analysis, year-based analyzes were made and it was monitored which year which country and which patent class directed to this field. At the same time, the network analysis was performed for the entire data set, and the country with the most patents and the class with the most patents in the field of information security of the last decade was determined. Patents are divided according to technical areas thanks to the classification system. In the time series analysis, approaching the frequency value from a different perspective, the most used nine patent class frequencies in the data set were used and the results of the time series analysis made in this way were shared. Different methods were tried for time series analysis, and the estimates made for 2020 by the method that made the closest estimate for 2019 were taken into account. Our evaluation reveals that the proposed time series analysis method can calculate the number of patents belonging to patent classes of patents received in the field of information security for the first half of 2019 with 98% accuracy.