Research On Measure And Application Of Normal Cloud Similarity

Because there are many uncertain information in the real world,reasonable description and processing of uncertain information is the key.As a theoretical model for dealing with uncertain information,cloud model can realize the uncertainty transformation between qualitative concepts and quantitative descriptions.Many scholars have applied cloud model to decision analysis,data mining,intelligent control and other fields.As an important part of cloud model research,similarity measurement methods between cloud concepts play an important role in various fields of application.Since the existing cloud concept similarity measurement methods mainly consider the cloud concept digital characteristics and the overlapping area of cloud concept characteristic curves,the measurement results are one-sided and the ability to distinguish is weak.Therefore,the main goal of this thesis is to study a more effective cloud concept similarity measurement method.Based on this,the main research work of this thesis includes the following three aspects:（1）A concept similarity measurement method of normal cloud based on Hellinger distance is proposed.Firstly,the similarity between two probability distributions is characterized by Hellinger distance,the Hellinger distance between two normal distributions is given,and its mathematical properties are discussed.Secondly,taking the characteristic curve of the cloud model（such as expectation curve,internal and external envelope curve,etc.）as the entry point,integrating the distribution characteristics of the normal cloud concept,a Hellinger distance based on the normalized characteristic curve is constructed on the basis of the Hellinger distance between two normal distributions,so as to realize the similarity calculation between the two normal cloud concepts.This measurement method not only considers the digital characteristics of the cloud concept,It also takes into account the distribution characteristics of the normal cloud concept.（2）Based on the Hellinger distance measurement method of normal cloud concept proposed,two similarity algorithms of normal cloud concept are designed,and experimental simulation and application analysis are carried out.First of all,through simulation examples and classification experiments on UCI time series data sets（synthetic control chart data sets）,the existing similarity calculation methods and the methods proposed in this thesis are compared and analyzed for concept difference and classification error rate.The experimental results show that this method has higher concept difference and lower classification error rate.Secondly,the proposed algorithm is applied to the collaborative filtering recommendation system,and the experiment is carried out on Movie Lens100k film review dataset.The average absolute error（MAE）and root mean square error（RMSE）are used to compare the algorithm in this thesis with LICM,ECM,MCM and other classical methods.The experimental results show that the method in this thesis has low recommendation error,good classification ability and recommendation effect.（3）This thesis uses the similarity measurement method of normal cloud concept proposed in this thesis to comprehensively evaluate the air quality of Yinchuan City.Firstly,the mass concentrations of six major atmospheric pollutants,SO₂,NO₂,PM₁₀,PM_2.5,CO and O₃,are selected as evaluation indicators,and the numerical characteristics of the standard cloud concept are calculated according to the concentration limit standards of these six pollutants in the Ambient Air Quality Standards（GB3095-2012）;Secondly,according to the historical atmospheric data released by China National Environmental Monitoring Station,the concept of cloud is generated from the data of six pollutants in Yinchuan City from 2019 to 2021 through reverse cloud transformation;Finally,the similarity between the actually generated cloud concept and the standard cloud concept is calculated by the cloud concept similarity measurement method proposed in this thesis,and the evaluation grade is divided according to the similarity and the evaluation results are analyzed.The analysis results show that the main air pollutants in Yinchuan City are PM_2.5 and PM₁₀,and the air quality is generally stable and good.