Research And Application Of Statistical Data About Human Resource Exception Detection Based On Data Mining

With the development of national economy, department of statistics requests more and more to the data reported by enterprise. Some enterprise hides real data and tells a lie for their own benefit. How to distinct unreal data from the whole data becomes an urgent task in the field of statistics. Using data mining to detect abnormal data is a feasible method. But it hasn’t been widely used in this field. This paper does some research on methods and application of exception detection based on data mining. This paper’s work is listed as below.First, this paper introduces the technology of data mining and exception detection, and lays a lot of emphasis on K-means clustering algorithm and DBSCAN clustering algorithm.Then, this paper suggests an adaptive method of exception detection based on data mining. This method uses the improved K-means to get parameter K automatically and combines K-means algorithm with DBSCAN algorithm and sums up their advantages:It brings about self-adaptive of parameters and avoids artificial errors. Then the method clusters the date by the improved K-means and uses the result to compute DBSCAN parameters which will get abnormal data.In the end, this paper develops exception detection system based on the method which has been suggested. This system collects data and does data preprocessing and detects abnormal data. In order to solve the problem of data source, this paper develops another system which is called human resource and salary information system.What the paper has done solves the problems of hiding and lying in the field of statistics effectively and provides a new track to detect abnormal data in the field of statistics.