The Application Of Time Series Analysis In Educational Research

Education has become a common research by many scholars, teachers’ demand and the scale of development of higher education not only work on the teachers’arrangement and construction, but also affects the quality and quantity of its students.In this paper, Based on the study of no-stationary time series forecasting, the teacher’s demand and the scale of higher education are predicted. On account of no-stationary of the sequence, the data has been processed by two methods. The first method is through taking the logarithm, an Exponential Regression-ARMA model is established for the number of students. The second method is the data are decomposed and reconstructed by wavelet analysis, Then the reconstructed part is predicted by LS-SVM model, the detail part is predicted by ARMA model, finally the reconstructed part and the detail part are cumulated to predict the future of the scale of higher education. Better predicting performance can be obtained. Financial data show that the predicting method is effective.