Text Mining Based GPT Method for Analyzing Research Trends
摘要
In recent years, keywords are often extracted using text mining techniques when analyzing research trends to start a new study. In order to check whether the extracted keywords are relevant to the sentences, it is necessary to analyze the accuracy by using patterns, trends, etc. when extracting. In this study, we extract keywords using text mining techniques for a specific domain and compare and analyze the sentences generated by using the extracted keywords as input to GPT fine-tuned with data from the domain. By comparing the keywords extracted using text mining techniques with the sentences generated using GPT, it is verified that the results are relevant to the domain data. Through this process, it is expected that more accurate sentences can be generated by using the keywords extracted using text mining techniques as the result of GPT, unlike when only existing text mining techniques are used.