Based On Clustering Analysis Of The Outlier Detection Research And Its Application In The Audit

Since 90's in the last Century, the computer, the network and the communication technique have been developed rapidly; they have been expanded unceasingly in the fields of application, and also their influence and vitality have been permeated to the various aspects of society. What's the root of the challenge of AUDIT owing to information is that the field of economic management has been informationized. At the present time, the economic activities of state organs, Institutional and Corporate state have been achieved informatization, especially, with the development of the Internet, Electronic Commerce and Internet transactions producing, space limitation of financial management has been breakthrough, the development of Computer Audit is the trend of network, so the result is that financial payment of different agencies and online audit have been implemented.The working of audit department must be in keeping with huge change of audit object and audit environment. What's the corollary of development of audit theory and practice growing is that computer audit, which owns incomparable dominance contrasting with manual audit. Therefore, establish computer network audit system can be greatly improved the quality and efficiency of audit.Concrete the daily audit work, informatization brought maqnanirmity business data。The auditors often overwhelming, when they face it. Personnel distribution nor purposeful and planning. To solve these problems, we introduce the data mining technology. Currently, data mining as the assisted audit tools greatly improves the auditors to mass data control ability.This paper introduces the data mining, clustering analysis, outlier analysis and computer networking audit techniques, expounds the concept of several technical implementation of algorithm, etc. Then the key is discussed based on clustering analysis and data analysis of isolated points, the third is to describe the DBSCAN algorithm again and design. Finally a state-owned enterprise business data for mining object. Do the research of clustering analysis and outlier analyzed application.