Abstract: Examination is an important means of evaluating and supervising the learning and achievements of students in higher education. Computer-aided evaluation (CAE) system has been widely used in specialized courses teaching, such as biochemistry, genetics and other biology-related courses at the foundation stage of medical colleges and universities. It has significantly improved the efficiency of subject teaching while reducing the cost of examination organization and implementation. However, the conventional CAE platform basically provides support for the standardization of test questions in the form of objective assessment, while it has been difficult to standardize the evaluation of subjective test questions. In this study, we try to use string editing distance as the main reference index for text similarity comparison, and apply the data processing principle of artificial intelligence to text discrimination and reference scoring of subjective questions such as fillingin questions, noun interpretation, short answer questions and question-and-answer questions, and explore the corresponding evaluation strategies and weight rules for different types of questions. In addition, in the process of testing, we have also realized the integration of all hardware requirements of mobile terminals represented by cellphones, which greatly improved the applicability and flexibility of the original online examination system, and fully adapted to the formative evaluation or completion examination needs in different classroom scales and other scenarios. We conclude that the solution for grading subjective test using artificial intelligence can be feasible under well-organized examination with properly arranged answer keys, where string similarity distances can be used as the primary index for calculations and down-dimension of cognition levels should be the strategy for text extraction. The continuous development and optimization of our platform or system of such types will benefit the students’ learning by significantly reduce both the workload and operational errors of teaching staffs. 

Key words: mobile online examination platform, editing distance, subjective questions