Quantum Algorithm And Its Application In Data Mining

Along with the computer database technology and network technology unceasing development, the development of the information society, people for knowledge discovery and information processing efficiency needs become increasingly demanding, in this big background, data mining technology and quantum computing both were born.On the one hand, the data mining technology in the face the great contradiction cases of information expansion and lack of knowledge, constantly play the role of savior. Data mining’ research upsurge make it widely used in information acquisition, decision support, etc. And take the great responsibility of building Noah’s ark for dealing with the flood of information in the area of electronic commerce, scientific research, social networks, etc.On the other hand, the new information challenges of mass data, unstructured data, high dimensional data, missing data, noise data, distributed, etc. become more and more serious. Data mining technology who in the top of the wave is also become at their wits’ end. But quantum computation and quantum algorithm has unique advantages in solving complex problems.It can even change the traditional world helpless NP problem into P problem, which making the problem solved effectively.Therefore, this paper tries to prospective research existing quantum algorithms, analyzes its advantages and disadvantages, puts forward or improve new quantum algorithm, and try to fusion quantum algorithm and the data mining algorithm, for new information challenge to seek a way out. This paper mainly does the following several aspects work:Firstly, this paper introduces the quantum computing and data mining related background, makes a brief review of the basic principle of quantum computing, and discussion and the analysis the quantum walk theory.Secondly, we put forward a3D angle coding quantum genetic algorithm (3D-AQGA).This algorithm makes full use of quantum space movement characteristics, using the polar coordinate Angle coding way as the starting point, the algorithm’s updating and variation operation was redesigned. Simulation results also show that the algorithm in optimizing problem has obvious advantages.Thirdly, considering the importance of the distance metric in clustering problems, we proposed quantum genetic clustering algorithm based on the previous quantum genetic algorithm. Combined with3D-AQGA and traditional k-means And full consideration of data set attribute correlation, dimension, noise, etc., design a kind of based on quantize range of the generalized weighted Minkovski distance, and alternative Euclidian distance as this chapter clustering algorithm’s distance measure, experiments show that the new algorithm for clustering effect has improved significantly.Fourthly, we proposed a grid quantum walk clustering model through a deeply research on quantum walks. Considering the unique characteristics of clustering model, we change existing discrete quantum walk model using grid and analyze and discuss the new model. Algorithm can effectively complete clustering task and have a exponent speedup.