what are toefl reading comprehension passages about? a topic modelling approach

Toefl reading passages cover a wide variety of different subjects that are highly technical and are written in an academic style. therefore, helping toefl test takers to acquire background knowledge of the frequently-used topics of toefl exam, not only facilitates answering the questions of the reading section but also provides insights for better performance in writing and speaking sections. topic extraction could be fulfilled with topic modelling (tp) which is a novel approach that uses a collection of machine learning techniques that automatically identify the thematic information of a collection of texts. tp offers new ways to automatically extract semantic domains without imposing pre-defined categories on the data. it also provides quantitative results in the form of lists of words and phrases that indicate common thematic patterns that can be examined qualitatively to clarify aspects of discourse. this study attempts to reveal the most common themes of toefl reading passages. to this end, a corpus of roughly 85000 words from nine toefl books was compiled. to run tp, machine learning for language toolkit (mallet) developed by (mccallum, 2002) was used. this process included two phases: first, the quantitative phases in which the software was instructed to extract topics; second, the qualitative phase in which each topic was interpreted based on the researcher's inspection of each list. the findings of this study would be benefical for toefl instructors and test takers to choose appropriate reading materials.