Research on English Text Keyword Generation Based on Machine Learning

With the advent of the “global village” era, the changes in domestic and foreign affairs have become increasingly frequent. As a result, there will be many English texts such as news newspapers, journals, magazines, dialogues, comments, etc. expressed in English. Among them, some English texts are even lengthy and difficult to understand quickly. In order to provide people with a quick way to grasp key information in English texts, this paper proposes a machine learning based research method for generating English text keywords. The first stage is to collect some English text data for analysis. This study randomly obtained some movie review text data online as the research text; The second stage is to learn and generate keyword information based on the principles of TDF-IDF and TEXTRANK machine learning algorithms (set to 5 in this article’s test), while separating the corresponding keywords, weight coefficients, and part of speech information and presenting them in a centralized comparative drawing. According to the research and testing in this article, it can be concluded that through the calculation of TDF-IDF and TEXTRANK algorithm models, any text can quickly extract the desired keyword information and has universality.