The Research Of Implicit Learning Which Used Time Series Analysis Method

Along with the development of Psychology, longitudinal methods become popular, and the time series analysis method is one of the most important methods of them. Although the theory is very complicated, and needs a long process to get data, it has very significant application value to study psychology problems. The method not only can compare the tendency among variables, make sure the cause and effect relation between dynamic variables, but also can forecast and control human's mentality and behavior.Based on the overviews of former researches, the study chooses implicit learning in the field of cognition and learning as object, uses the idea of time series to design experiment and work out the test software to get data, and then use the time series analysis method to analysis data. The targets of this study are to apply the method to the field of psychology, study the inner mechanism and rules of implicit series learning and forecast the rules. The main results of this study are as follows:1. The series learning is exactly existent, participants can't realize the rules of series, so the series learning is implicit, but not explicit.2. Time series analysis shows that the mechanism of implicit series learning is horizontal association of events, not vertical association. They learn the adjacent association of stimulate position, confined by physical forms, and don't produce abstract knowledge.3. Under the three RSI conditions, we use the method of ARIMA to find that rules of implicit series learning separately fit the AR(2),AR(1) and ARMA(l,0,2) model, the results shows the forecasting effect is perfect.4. Using exponential smoothing methods to simulate the rules of the implicit series learning under the three RSI conditions, including single and double exponential smoothing procedure, the results shows the simulating effect is perfect.