Patent Prediction Based on Long Short-Term Memory Recurrent Neural Network

Abstract

Patent prediction is not only an efficient way to grasp development tendency of relevant science and technology fields but also has been widely used in the areas of patent recommendation, evaluation and transaction. However, with the rapid increase of patents, it is becoming more difficult to mine potential information from huge amount of patent records in database. Based on sample data analysis with long short-term memory recurrent neural network model, we propose a patent prediction scheme. The implementation process of the proposed scheme is discussed, and the examples to predict railway transportation patents in international patent database are given. The calculation results of root mean square errors show that our scheme can obtain higher prediction accuracies than traditional auto-regressive-moving-average model.