Interval Set Rough Set Theory And Clustering Algorithm Research

Machine learning is a principal method in the field of artificial intelligence.It can analyze disorganized data and gain the knowledge it contains.Clustering algorithm is a typical unsupervised learning in machine learning,which has been successfully applied in the real world.Traditional clustering algorithms,however,mostly consider different clusters as a division of the entire data set.But in practical applications,the relationship between data elements and clusters is not just simply attributed to a clustered situation.For this more complex situation,this paper designs a new algorithm model and applies it to the real recommendation scenario.This dissertation focuses on the theory of interval set rough sets and clustering algorithm.The interval set rough set,firstly,is defined.Then its structure and properties are analyzed and studied;the K-means of clustering density peaks with means and irregular shapes are studied.A clustering algorithm model for rough set machine learning in the granular computing framework is established,and the design and implementation of the interval set rough set clustering recommendation system is established.Based on the rough set and clustering algorithm in machine learning,this dissertation mainly does the following work:1.Describe the theory of interval set rough set in detail.Explore the rough set from the perspective of interval set.These theories combine interval sets and rough sets,using not only the interval sets of interval sets,but also the ideas represented by rough set approximate sets.The algebraic properties and basic structure of the model are studied,and the uncertainty measurement and attribute reduction are also studied.2.Based on the Interval Set Rough Set Clustering Algorithm(ILRSKM),K-means algorithm is improved by the interval set rough set,and divided the clusters up and down under the interval set rough set theory to make full use of information of clusters.The information of boundary uncertainty measurement is used and better clustering results are obtained.3.A Density-Peak Clustering Algorithm based on Interval Set Rough Sets(ILRSDPC)is proposed.The density set clustering based on interval set rough sets is studied,and clusters are clustered on the boundary of uncertainty under the theory of interval set rough sets.Better clustering results have been achieved.4.Design and implement an application system based on interval lattice rough set recommendation: The interval lattice rough set,K-means,and DPC clustering algorithm are applied in a distributed environment.A crawler algorithm,firstly,is used to build a recommended front-end website.And then the user-generated learning behavior data is collected,transmitted to the recommendation module through the network,and the clustering algorithm was applied for research in this recommended scenario.In this dissertation,Interval set rough set theory is applied to the improvement of clustering algorithm and studied in the practical application recommendation scenario.The research result has a promising application space.