Study On Technology Paradigm Transformation Based On Patent Text Mining

With the rapid development of technology, the competition between enterprises is becoming fiercer. The transformation of technology paradigm has a significant impact on the related industries, the analysis of the development trend of technology paradigm affects the formulation of enterprises decision making, determines whether the enterprise is able to occupy a favorable position in the future. Because of the shortage of social resources, the development direction of technology paradigm also influences the national strategies, it is important to study and foresee technology paradigm transformation. The patent contains abundant technology, business and product knowledge, reflecting the technological innovation activities. The textual data of patent which is of great value contains the main topic, innovation and application of patent. The development of text mining technology provides a powerful tool for the analysis of textual date. The development of technology has a certain life cycle, and the life cycle theory provides a new prospect for the study of technology paradigm transformation. The text mining technology is used as the tool, and the technology life cycle is used as the research angle of view to analyze the method and theory of technology paradigm transformation.After recognizing the technology paradigm transformation characteristics, researching the mechanism of technology paradigm transformation, analyzing the relationship between patent activities and technology paradigm transformation, a method is proposed to foresee the transformation of technology paradigm based on text mining from the perspective of technology life cycle. The method firstly pre process the patent text data, remove the punctuation mark and the stop word. Secondly, The LDA model is used to extract the theme of patent, produce the probability distribution of patent document on the themes and construct a vector space model. Thirdly, k-means algorithm is used to realize patent document clustering and a theme identification method is proposed. Finally, the technology paradigm of each patent cluster is determined and the "S" curve is drawn with the time as horizontal axis, accumulative amount as vertical axis to identify the characteristics of technology paradigm transformation and foresee the transformation of technology paradigm.In order to study the effectiveness of the technology paradigm transformation foresight method based on text mining, the mobile phone communication technology was taken as an example to conduct an empirical study. Derwent patent database was used as data source and a database is constructed to storage related data. Java is used as program language to realize text segmentation and remove stop words from text. JGibbLDA implemented by java is used to extract topic of each patent and get probability distribution of patent document on topics. Then, k-means algorithm is conducted to realize patent clustering. Finally, loglet lab2 helped to draw a picture of "S" curve and estimate relevant parameter. From the empirical study a conclusion is drawn that 2G technology paradigm is transforming to 3G and 4G technology paradigm. The practicability and effectiveness of the proposed method are verified by empirical studies.