Feature Learning of Patent Networks Using Tensor Decomposition

In the age of big data, the graph clustering algorithms are visual analytics can help the decision-makers to have a precise description of the relationships between data information in the technological areas and companies. Considering the patent as a source of science and technologies this paper presents a new road map for the patent landscape begins with searching and recognizing related text data in similar patent using a deep learning approach to build a model that can exhibit the relationships between text patent in the form of a graph and constructing large patent networks, to complete the last pass in a patent network project and owing to the complexities of the data structure involve the application of the basic tensor concept and their properties to perform automatic unsupervised learning without taking account of the noisy patent data, throw this process the visualization of graphs knowledge much time more powerful and helpful could prove very useful in improving the decision-making process within an organization.