Forecasting Model And Application Of Tandem Trend In China 's Coastal Tourism

The development of China’s coastal tourism is affected by natural, economic, policy and other factors. These factors are also influenced by the external factors, so that the coastal tourism system in China is complex and changeable, and it is difficult to determine the time series of China’s coastal tourism industry are linear or nonlinear. Secondly, since China began to carry out the time of marine statistics is relatively late, so that the total amount of data that can be applied to forecast is not much. Furthermore, China’s marine economic accounting system has also changed many times, so that the total quantity index that can be applied to the prediction of the coastal tourism trend is single, the statistical data is not consistent, the inequality of data is more obvious, so the overall data quality is not high. But the existing forecasting models do not account these characteristics of the coastal tourism system, not to say to find a uniform paradigm(single or combination forecasting model) that can be used to accurately fit the development time series of the coastal tourism industry, and explain the reasons for the change of time series.From the existing literature is not difficult to conclude that the qualitative forecasting methods and quantitative forecasting method s have their own advantages and disadvantages, but they are interrelated and complement each other, so just because the prediction error of a certain model is too large to drop it is not scientific. A more scientific approach is to combine different individual forecasting methods in order to make use of the information of each model to reduce the impact of environmental random factors in the single prediction model.It confirms to test with practical examples that the paper puts forward the series trend forecasting model can fully take into account the characteristics of "poor" data and "difference" data of coastal tourism in China, so that the linear and nonlinear information of time series of coastal tourism can be extracted effectively. Series trend forecasting model solve the problem that the traditional method can only predict the changes of part of coastal tourism. And to a certain extent, it has enriched the research of the combination forecasting method of coastal tourism industry.The main logical structure of this paper: the first chapter describes the background and significance of this paper, the structure of the paper and the possible innovation points of the paper. The second chapter reviews the results of coastal tourism and combination forecasting,, and proposed that there are research gaps in the research of the series and parallel combination forecasting model. The third chapter analyzes and compares the theoretical results of qualitative prediction and quantitative forecasting techniques. The fourth chapter analysis of the characteristics of China’s coastal tourism. On the basis of this, a series trend prediction model that can fully exploit the time series characteristics of China’s coastal tourism is proposed. The fifth chapter, select the single model(trend extrapolation, exponential smoothing, ARIMA, GM(1,1), BP network based on LM algorithm(LM-BPNN)) which is suitable for China’s coastal tourism industry. A new combination model is obtained through the combination of P ?f(L, N) ??.After the redundant filtering of the new combined model, by the way of parallel and series to combine the new combination model. Finding the performance of series trend forecasting model is better than that of parallel trend forecasting model. And then use the series trend forecasting model to forecast the value of the added value of China’s coastal tourism in 2015~2017. In the sixth chapter, the conclusion and prospect of this paper, and think it is necessary to further study “Construct comprehensive evaluation index system for evaluating the development trend of coastal tourism” and “How to determine the combined weight”.