Classification method of patent texts in security field

Abstract

The invention provides a classification method of patent texts in the security field. The invention provides a classification method of patent texts in the security field. The method comprises the steps of: 1, adding words and expressions occurring frequently into a using-stop word list in the text processing process to save the storage space and improve retrieval efficiency; 2. introducing the pre-trained Word2Vec model to solve the dimension disaster problem caused by the traditional method; 3, extracting text features by train a Long short Term memory (LSTM) classification model to classifypatent texts in that security field; 4, evaluating the classification result by using an accuracy rate and a ROC curve evaluation model. Experiments show that this method can classify the patent texts in the field of security, train and test 50,000 patent texts, and the accuracy of the test set reaches 93.48%.

Description

technical field [0001] The invention relates to a method for classifying patent texts in the security field. Background technique [0002] With the rapid development of information technology and knowledge economy, the number of patent applications in my country is increasing day by day. As an intangible asset, patent has huge commercial and research value, and has become an important indicator to measure the comprehensive strength of various countries. How to obtain cutting-edge and innovative achievements from patent texts, transform them into products, and realize industrialization has become the focus of research by experts and scholars. As a basic work, patent text classification plays an important role in patent retrieval, patent mining, and strategic decision-making. Therefore, patent text classification has very important research significance and research value. At present, there are few researches on patents in the field of security. Since patents in each field ...

AI Technical Summary

Problems solved by technology

For example, convolutional neural network performs text classification. When constructing text semantics, it is necessary to manually set a window to capture context information, and the size of the window has an important impact on the classification results, and it takes a lot of time during the training process; cyclic neural network When performing text classification, because the recurrent neural network has a deep memory of the last input signal and a shallow memory of the early input signal, this will lead to the problem of "gradient disappearance"

Images