高校招生数据分析与应用研究：基于LSTM技术的招生推荐系统 构建

lleges in China from 2007 to 2017, except for the National University of Defense Technology, in various provinces for arts and sciences majors. The original data was cleaned, missing values were filled, and group integration was carried out according to actual requirements, combined with feature engineering for exploratory data analysis. The trends in the number of majors offered by each college in recent years, the most competitive, potential, and least competitive, potential majors were statistically determined. In terms of models, this paper cleverly used batch file allocation scores and total scores of top students as control targets, and used deep learning algorithm LSTM (Long Short-Term Memory Neural Network) to predict the relative thresholds of each major for the next year. By using the Analytic Hierarchy Process model to compare all decision-making scenarios pairwise for candidates, a college entrance examination recommendation system was established; recommending corresponding colleges and majors. The prediction accuracy of LSTM is very high, with good practical significance, and the consistency test mechanism in the Analytic Hierarchy Process can minimize the loss of objectivity in the model caused by significant individual subjective comparison biases, making the overall model practical and scientific. Keywords: crawler; LSTM; Analytic Hierarchy Process; college admission; recommendation system.