Date of Submission
Fall 12-16-2020
Degree Type
Thesis
Degree Name
Master of Science in Computer Science (MSCS)
Department
Computer Science
Committee Chair/First Advisor
Dr. Selena He
Track
Big Data
Chair
Dr. Selena He
Committee Member
Dr. Meng Han
Committee Member
Dr. Xiaohua Xu
Abstract
Digital currency is a novel topic. Due to the publicity of the Internet and the investment value of the digital currency itself, many people hope to participate in the digital currency-related financial market. However, many people still hold a wait-and-see attitude at this stage. The main reason is that digital currency is a relatively new field of financial investment, and there is no scientific system and rigorous data to provide investors with sufficient information. In this paper, we designed a framework based on a natural language processing model to calculate the innovative capabilities of digital currency-related companies, which can help to understand the company's development trend and provide investors with objective data recommendations. First, we determined and selected the patent data as an indicator for calculating the innovation of technological development and established a relational database for extracting relevant information from patent data effectively. Second, we proposed a semantic level natural language learning model based on patent data to obtain the potential development value of related technology fields or companies. The model can effectively classify patent data, calculate the potential value of the patent, and summarize the innovation of the selected topic. Finally, we used all the patents from multiple digital currency-related technology companies as training data and verified the effectiveness of our designed model by referring to the performance of different machine learning models. At the same time, we also calculated the company's development potential and the market potential value of all patents owned by the company. As the experimental result shows, the framework we proposed can achieve a good effect and be suitable for calculating the innovation value of different technological fields or companies.