Research on Student Performance Prediction based on BKA-BiLSTM Algorithm

Based on the online resources and online data of Object Procedural Programming (OPP) course, this paper proposed the course dataset of students (Dataset of OPP: DOPP), and conducts a prediction study on students' final exam scores through this dataset. The hybrid algorithm of Black Kite Algorithm (BKA) and BiLSTM: BKA-BiLSTM algorithm model, is used for prediction, which can not only capture the data upstream and downward information, but also optimize the hyperparameters of the model through BKA. The results show that BKA-BiLSTM has better results in evaluation index than other regression analysis algorithms for the dataset DOPP of this course, and has good practical application value.