2024 ASEE Annual Conference & Exposition

Many non-English speaking international students come to the United States to pursue undergraduate engineering programs. However, most of them struggle to learn and use English proficiently. This struggle to learn and use English poses various challenges. For example, such students struggle to describe their plans and thoughts to their college peers and colleagues at work. Also, it is mostly harder for such students to make their place in academic or industry careers. Some of these difficulties arise because students cannot identify sentence structures or differences between various types of sentences in English. Writing in complete sentences is one way to convey ideas effectively in English, and this paper presents the model and its accuracy results for the different types of English language sentences. These types include declarative, imperative, interrogative, exclamative, or invalid. We hypothesize that this model will help students classify written sentences as declarative, interrogative, imperative, exclamative, or invalid. We also discuss the future applications of this model and believe that it can help engineering students correct sentence structure errors according to sentence types. We considered 100 sentences of each sentence type for accuracy and calculated various measures, including precision, F1 score, and recall. Out of 100 declarative sentences, 92 were properly identified as declarative sentences, scoring a high accuracy score of 92%, a precision of 95.8%, a recall of 92%, and an F1 score of 93.9%. Out of 100 interrogative sentences, 77 were correctly classified as interrogative sentences, scoring a moderately high accuracy score of 77%, a precision of 95%, a recall of 92%, and an F1 score of 95.5%. Out of 100 imperative sentences, 55 were correctly classified as imperative sentences, scoring a lackluster accuracy of 55%, a precision of 98.2%, a recall of 55%, and an F1 score of 70.5%. Lastly, out of 100 invalid sentences, 81 were properly determined as invalid, scoring a moderately high accuracy score of 81%, a precision of 50.6%, a recall of 81%, and an F1 score of 62.3%.