The rapid pace of technological advances creates many difficulties for R&D practitioners in analyzing emerging technologies. Patent information analysis is an effective tool in this situation. Conventional patent information analysis has focused on the extraction of vacant, promising, or core technologies and the monitoring of technological trends. From a technology management perspective, the ultimate purpose of R&D is technology commercialization. The core of technology commercialization is the technology transfer phase. Although a great number of patents are filed, publicized, and registered every year, many commercially relevant patents are filtered through registration processes that examine novelty, creativity, and industrial applicability. Despite the efforts of these selection processes, the number of patents being transferred is low when compared with total annual patent registrations. To deal with this problem, this study proposes a predictive model for technology transfer using patent analysis. In the predictive model, patent analysis is conducted to reveal the quantitative relations between technology transfer and a range of variables included in the patent data.

In this study, a novel three-dimensional (3D) bone morphogenic protein-2 (BMP-2)-delivering tannylated polycaprolactone (PCL) (BMP-2/tannic acid (TA)/PCL) scaffold with anti-oxidant, anti-inflammatory, and osteogenic activities was fabricated via simple surface coating with TA, followed by the immobilization of BMP-2 on the TA-coated PCL scaffold. The BMP-2/TA/PCL scaffold showed controlled and sustained BMP-2 release. It effectively scavenged reactive oxygen species (ROS) in cells, and increased the proliferation of MC3T3-E1 cells pre-treated with hydrogen peroxide (H₂O₂). Additionally, the BMP-2/TA/PCL scaffold significantly suppressed the mRNA levels of pro-inflammatory cytokines, including matrix metalloproteinases-3 (MMP-3), cyclooxygenase-2 (COX-2), interleukin-6 (IL-6), and tumor necrosis factor-α (TNF-α), in lipopolysaccharide (LPS)-induced MC3T3-E1 cells. Furthermore, it showed outstanding enhancement of the osteogenic activity of MC3T3-E1 cells through increased alkaline phosphatase (ALP) activity and calcium deposition. Our findings demonstrated that the BMP-2/TA/PCL scaffold plays an important role in scavenging ROS, suppressing inflammatory response, and enhancing the osteogenic differentiation of cells.

Many studies have predicted the future performance of companies for the purpose of making investment decisions. Most of these are based on the qualitative judgments of experts in related industries, who consider various financial and firm performance information. With recent developments in data processing technology, studies have started to use machine learning techniques to predict corporate performance. For example, deep neural network-based prediction models are again attracting attention, and are now widely used in constructing prediction and classification models. In this study, we propose a deep neural network-based corporate performance prediction model that uses a company’s financial and patent indicators as predictors. The proposed model includes an unsupervised learning phase and a fine-tuning phase. The learning phase uses a restricted Boltzmann machine. The fine-tuning phase uses a backpropagation algorithm and a relatively up-to-date training data set that reflects the latest trends in the relationship between predictors and corporate performance.

In today’s knowledge-based society, industry-university cooperation (IUC) is recognized as an effective tool for technological innovation. Many studies have shown that selecting the right partner is essential to the success of the IUC. Although there have been a lot of studies on the criteria for selecting a suitable partner for IUC or strategic alliances, there has been a problem of making decisions depending on the qualitative judgment of experts or staff. While related works using patent analysis enabled the quantitative analysis and comparison of potential research partners, they overlooked the fact that there are several sub-technologies in one specific technology domain and that the applicant’s research concentration and competency are not the same for every sub-technology. This study suggests a systematic methodology that combines the Latent Dirichlet Allocation (LDA) topic model and the clustering algorithm in order to classify the sub-technology categories of a particular technology domain, and identifies the best college partners in each category. In addition, a similar-patent density (SPD) index was proposed and utilized for an objective comparison of potential university partners. In order to investigate the practical applicability of the proposed methodology, we conducted experiments using real patent data on the electric vehicle domain obtained from the Korean Intellectual Property Office. As a result, we identified 10 research and development sectors wherein Hyundai Motor Company (HMC) focuses using LDA and clustering. The universities with the highest values of SPD for each sector were chosen to be the most suitable partners of HMC for collaborative research.