Study On Optimization Of Quantum Search Algorithm And Its Applications In Business Intelligence

Business Intelligence(BI)is the advanced stage of enterprise informatization.It focuses on transforming business data into knowledge,so as to enhance the competitive advantage of enterprises.BI collects and manages business data through information technology,then uses artificial intelligence and machine learning to mine the data,and finally obtains the strategic knowledge to serve the decision-making.With the development of information technology and data collecting techniques,the volume of business data is growing explosively,so that BI can not complete the task of decision support in time.Quantum computing achieves tasks utilizing the characteristics of quantum superposition and entanglement.Compared with classical computing,quantum computing has an obvious acceleration effect in solving some specific problems.For example,Shor’s algorithm has exponential acceleration over its classical counterpart in factoring large prime integers;Grover algortihm has a quadratic speedup over classical search algorithms in searching disordered databases.In recent years,quantum computing has also been gradually applied to the field of BI,and a number of efficient quantized BI technologies have been proposed.At present,quantized BI technology is an innovative research direction with high academic value.However,the research of quantized BI is still in the development stage,and there are many BI technologies without corresponding quantum algorithms.In this paper,some BI technologies without efficient quantum algorithms are studied.Based on the improvement of quantum search algorithm,some quantum BI technologies are proposed.The proposed technologies are expected to play some role in the upcoming quantum era.In general,this study includes the following aspects:(1)Several adjustments of quantum search algorithmQuantum search algorithm is a sharp tool to solve the search problems in BI.However,the existing algorithms have some defects.In order to better combine it with BI,several adjustments of quantum search algorithm are proposed.Firstly,this paper proposes a multi-operator quantum search algorithm and a phase-self-adaptive quantum search algorithm.When the target proportion is known,the two algorithms can always obtain the target component with 100% probability.Secondly,this paper proposes a segmented quantum search algorithm,which can maintain a success probability of more than 93% when the value range is known.Finally,this paper proposes an improved version of Grover auto-control algorithm.The algorithm ensures the success probability of 85.36%.(2)Association rule mining based on partially quantized rough setAttribute reduction is an important preprocessing technology in association rule mining,and the computation of Core is an important step of attribute reduction.The efficiency of the computation of Core is related to the efficiency of attribute reduction,and then affects the timeliness of association rule mining.In order to improve the timeliness of association rule mining,a quantum algorithm for the computation of Core is proposed and applied to association rule mining based on rough set.(3)Non-parameter outlier detection based on quantum searchOutlier detection technology is an important technology in BI.It can be used for business fraud detection,user behavior anomaly detection and so on.However,the existing outlier detection algorithms can not effectively deal with the inconsistent-sparsed and huge-volumed business data.In order to improve the efficiency of detection,this paper first presents an outlier detection algorithm based on natural neighbor graph;Then,by constructing two distance quantum black boxes,the quantization of natural neighbor graph is realized.Theoretical analysis proves the advantage of the algorithm in complexity;Data experiments verify the improvement compared with other outlier detection algorithms.(4)Customer segmentation based on quantized Na NG-DBSCANClustering is an important task in customer segmentation.However,there are some problems in customer relationship data,such as inconsistent data density problem,irregular data mode problem,parameter selection problem and so on.In order to solve these problems,this paper first proposes an Na NG-DBSCAN clustering algorithm based on quantum search;Then,a customer segmentation model based on quantized Na NG-DBSCAN is constructed.Data experiments verify the effectiveness of the model.