Use of Machine Learning at the Patent office to Track Global Trends in

Abstract

The present day world comprises several complex systems. These intricate systems require complex and sophisticated tools for their analysis in order to make precise forecasts about them which might enhance societal well-being. Use of modern day disruptive technologies based on artificial intelligence can be used to comprehend many of these intricate systems which remain unexplored by professionals. Accurate machine-learning algorithms have potential use in several disciplines such as medicine, climate change, traffic patterns, and criminal recidivism. This paper endeavours to leverage machine learning for analysis of patent data in order to facilitate innovation in the public health sector. The research focussed on the objective of developing computational methods to classify and cluster patent documents. Thus, aiding policy makers, legal experts and the research community in navigating the rapidly evolving landscape of patented healthcare technology. A three-phase approach methodology was adopted to meet the objective of the research. The first phase comprised data collection and processing, the second phase focussed on machine learning model development and the final phase included application and validation of the developed tools. The study became relevant in the wake of the fourth industrial revolution and the Covid Pandemic which has resulted in enhancing the significance of the healthcare industry exponentially. This exponential growth has resurfaced certain pertinent issues at the interface of intellectual property regime and the public healthcare system. Particularly the implications of the patent system on public health are highly debated. It becomes very crucial, more so in the case of a developing nation, to study the effect of the patent system on innovation, technology transfer and industrial dynamics with reference to the pharmaceutical sector. This study will be crucial in addressing some of these issues which concludes with development of ML based patent analytics tool and suggests that such an investment in machine learning based patent analytics tool helps to reduce patent prosecution and litigation thus providing a cost effective solution.