Abstract:

One of the major problem in improving of educational system is the problem of educational knowledge discovery about the future trends and numbers of student enrollments. This knowledge shows the level of students interests and some others factors, such as standardized test scores, economic data, competitor institutions, ets. This problem is unique for each educational system (institutional, school/college‐level, department‐level) and accurate enrollment predictions are important for budget, programs and personnel planning. To solve this problem time series forecasting techniques and models can be applied. There are two approaches to time series forecasting and analysis. The first one uses some hypothesis about dependences between time series. However often these dependences are unknown and there are not enough data about student enrollments to put out and test the hypothesis (typically, educational time series are short-term and have 7- 40 time points). Therefore in our research we apply models of the second approach, based on forecasting time series one by one. This approach is widely spread for time series with different lengths; many methods and forecasting models have been proposed. Some of them, statistical models are optimum for a wide class of time series. However, they frequently inadequate in forecasting short-term time series of student enrollments. To make short-term time series forecasting more adequate and accurate the fuzzy time series models were presented. The aim of this contribution is to show the opportunities of applying of fuzzy time series models to forecast enrollments and their trends. In this paper the enrollments forecasting software, based on the set of fuzzy techniques, is proposed. In the conclusion the received results are discussed and the efficiency of the proposed approach is shown.

Keywords:

Educational system, enrollments, forecasting, fuzzy models, trends.