Emerging Technology Identification Algorithm Of Study Based On Citation Analysis And Deep Learning

Emerging technologies have a deep impact on the development of enterprises,It guide the enterprise to keep up with technological develop trends. Knowledge of emerging technological develop trends can help enterprises make important tech-nical and business decisions. Therefore, it is very important to forecast emerging technologies yet most of the existing models identified for certain specific techni-cal fields usually have some limitations and can't meet the actual requirements.In view of this, this study focused on the following :First, with an in-depth analysis of the Patent citation data between 1975 and 2009, and extract some representative features of emerging technology identifi-cation. According to the above, extracted all the features of the patent citation data for the indexing and clustering and then applied the proposed emerging technology and non-emerging technology category automatic labeling algorithm,The experimental results show that the automatic labeling algorithm achieves good results. Then, based on the depth belief network and the logistic regres-sion model, this study constructed the emerging technology recognition algorithm based on deep learning.Secondly, used the patent citation data provided by the USPTO to test and evaluate emerging technology identification algorithms based on deep learning.The algorithm identifies the emerging technology by automatically selecting the best combination of features in the patent citation data. The results show that the proposed method can accurately and stably identify emerging technologies.Compared with other algorithms, the accuracy of emerging technology recognition algorithm is found to be 1.05% higher.