The evolution of patent mining: Applying bibliometrics analysis and keyword network analysis Farshad Madani * , Charles Weber Portland State University, USA article info Article history: Received 16 June 2015 Received in revised form 12 May 2016 Accepted 16 May 2016 Keywords: Technology mining Patent mining Bibliometrics analysis Keyword network analysis Cluster analysis CiteSpace abstract Text mining methods allow researchers to investigate technical documents (tech mining) and specifically explore patents for valuable information (patent mining. To the review literature and analyze the evo- lution of patent analysis and patent mining methods, bibliometrics analysis and keyword-based network analysis is applied on 143 papers extracted from the 'Web of science' database. Bibliometrics analysis was applied to determine top players researching in patent mining. Applying cluster analysis on the keyword network shows three main stages of patent analysis evolution. Also, it is discussed how patent mining is evolutionized in terms of information retrieval, pattern recognition and pattern analysis. © 2016 Elsevier Ltd. All rights reserved. 1. Introduction Patent databases are an important source of information for innovators [1e3],R&D engineers [4,5], corporate executives [6e9], and policy makers in technology latecomer countries [10e12]. In- novators need information on prior art, in order to assess whether their inventions are commercially viable [13].R&D engineers, who are trying to solve a particular technical problem, want to identify patents that may contribute to the solution of their problem [4,5]. Corporate executives, who are looking for a technology that fits their product strategy, will make use of patent searches to identify how and where they can gain access to the desired technology [6e9]. Policy makers in technology latecomer counties tend to conduct patent analyses, in order to identify particular gaps in the capabilities of their national innovation systems [10e12]. In all the above instances, patent databases serve as a critical source of in- formation upon which policy decisions are based. For patent databases to be helpful in decision making, the in- formation that they provide must be accurate, presented in a comprehensible format and delivered in a timely manner. This can only be done if the users of patent databases have access to capabilities in keyword extraction, pattern recognition and pattern analysis. These crucial aspects of modern text mining have thus become an integral component of decision making, both at the strategic and tactical levels. Patent citation analysis and even established statistical tech- niques like Term Frequency-Inverse Document Frequency (TF-IDF) for patent keyword analysis do not provide the user of patent da- tabases with an understanding of the content and the context of the patent. The user cannot determine whether a patent contains relevant prior art, unless he/she actually reads the patent. This process is highly inefficient. Tens of millions of patents reside in the databases of the world's major patent offices. The innovator may take years to identify all patents that are relevant. In order for patent databases to be useful for the abovementioned stakeholders, the processes for extracting and analyzing relevant information must be highly efficient. Researchers in academia have made sig- nificant progress in the area of applying text mining for keyword extraction [14] and pattern recognition [15e18]. However, the field of pattern analysis is still in its infancy by comparison. Due to advances in natural language processing, text mining methods and tools have become increasingly available in many different research areas including technology management where scholars try to extract useful information and textual patterns from technical documents, particularly patents. Applying text mining methods in technical documents is named ‘tech mining’ or * Corresponding author. E-mail addresses: fmadani@pdx.edu, farshad.madani@gmail.com (F. Madani). Contents lists available at ScienceDirect World Patent Information journal homepage: www.elsevier.com/locate/worpatin http://dx.doi.org/10.1016/j.wpi.2016.05.008 0172-2190/© 2016 Elsevier Ltd. All rights reserved. World Patent Information 46 (2016) 32e48