Classification Algorithm Based On The Interval Valued Fuzzy Sets

Classification is an important research topic in data mining and pattern recognition ,in many fields it has a wide range of application. At present, the classification algorithm mainly has the following questions: classification with the unknown process, the classification method is too complicated and difficult to achieve, classification results are poor interpretative.In view of the above questions, this paper studies classification algorithm based on the interval-valued fuzzy sets, especially for high-dimensional data, how to get better classification results.First of all, this article analyzes the interval-valued fuzzy sets and fuzzy sets with their related concepts and basic nature, then analyzed and discussed the distance between interval-valued fuzzy sets, as well as the the calculation of similarity.After analysis the shortcomings of traditional fuzzy inference rules, introduce interval-valued fuzzy production rules, then a fuzzy inference rules will be transformed into an interval-valued fuzzy sets, then calculating two fuzzy inference rules similar will be convert to calculate two interval-valued fuzzy sets similar.In this paper, a classification algorithms based on weighted interval-valued fuzzy set is bringed forward, this method takes into account the classification contribution of each property ,and through simulation to validate the effectiveness of the method.Second, for high-dimensional data, Fuzzy classification algorithms are relatively less, with the data dimension increases, interval-valued fuzzy sets are also on the rise of the discourse range, which not only increases the complexity of the calculation, which has also affected the results accuracy of the classification .In view of this situation, this paper improve the original weighted matching algorithm based on interval-valued fuzzy sets, and introduce the radar to reduce the data dimension, high dimensional data is expressed to the radar and use the strategy of holographic characteristics to reduce the data dimension, then use the improved classification method which based on the interval-valued fuzzy sets to identify different types of radar graphics, then through simulation to validate the effectiveness of the method.