Study On Methods Of Foreign Trade Volume Forecasting Based On Soft Set Theory

As an important part of China's economic system, foreign trade plays a key role in economy developing, and it is one of the troikas of China's economic growth. With the development of foreign trade, the dependency on foreign trade of China is higher. Therefore, the healthy development of China's foreign trade is a vital aspect of maintaining the economy security of China. It is necessary to strengthen the forecasting and monitoring the foreign trade volume, and to establish some forecasting and monitoring mechanism to prevent trade risk, so as to reduce the loss of our foreign trade and realize the sustainable development of economy in China. At present, the methods on foreign trade volume forecasting are mainly the bilateral trade with gravity model, qualitative analysis, univariate time series analysis etc. These information which these methods consider are certain-quantitative-hard information. While the studies of recent years show that soft information such as trade risks and policies etc. has more and more influence on China's foreign trade. Therefore, traditional forecasting methods that consider only hard information cannot be applied to solving these problems. However, methods that consider multi-information combining are absent. Thus, this dissertation employs soft set theory, a emerging mathematical tool for uncertain problem, and considers soft-qualitative information that influence foreign trade to study soft-set-based multi-information combining foreign trade volume forecasting methods.Firstly, individual forecasting methods are descripting the actual data by accuracy, which is fuzzy like high or low. And the key of combined forecasting method is how to choose optimal weight of each individual forecasting method. Thus, this dissertation proposes a combined forecasting method based on fuzzy soft sets. To validate the model, this dissertation reviews former literatures and makes some comparison, and uses some real foreign trade time series in experiments. The results show that this method improves the accuracy of forecasting, it has lower computational complexity and higher robust with fewer restriction. It can satisfy the requirement of forecasting foreign trade volume for government.Secondly, this dissertation proposes the concept of bijective soft set and defines some operations of it. Moreover, it analyzes the properties of it and some meaning of it on management area and potential applications as well. Furthermore, this dissertation proposes a foreign trade volume forecasting model based on bijective soft set, which considers soft information that influence foreign trade volume and uses the reductions of bijective soft set decision system to induce forecasting rules, and proposes an example. Additionally, it considers the situation that recent data has more prior to propose corresponding model and example. At last, it compares traditional methods with proposed methods. The results show that traditional methods that considers only hard information cannot be applied to these situation and the proposed methods can forecasting under multi-information combining situation. They have minimum requirement on samples, fewer restrictions, and get vivid-convenient forecasting rules.At last, based on the concept of bijective soft set, this dissertation considers exclusive-disjunctive data to propose the concept of exclusive disjunctive soft set, defines some operations of it and analyzes the properties of it. Moreover, this dissertation considers the situation that soft information for foreign trade forecasting includes exclusive-disjunctive data, to propose a foreign trade forecasting methods based on exclusive disjunctive soft set. It gives an example and corresponding algorithm. The example shows that exclusive disjunctive soft sets based method considers additional exclusive disjunctive data compare with bijective soft set based method. It extended the applicable scope of model, which can solve problems that include exclusive disjunctive data. The proposed method can obtain forecasting method and predict foreign trade volume by them.