Research Of The Minimum Living Security Decision Support System Based On The Data Mining

With the standardization of rescue operations and the improvement of operating maturity at all levels in civil affairs, the departments accumulated a large number of minimum living raw data over these years. However, all the various rescue operations data for different business objectives relatively independently stored in the different databases, which can not be shared with each other. So the decision makers were unable to extract valuable information from them. How about data acquisition, processing and effectively use of these heterogeneous, different protocol data become a big problem urgently to be solved in the civil affairs departments. Data warehouse technology and data mining technology are developed related to the new technology of database and artificial intelligence in recent years.They can be used to complete wide-scale analysis from various angles and find the inherent law based on a large amount of data operations such as classification, clustering and correlation analysis.Thus the data mining technology, data warehouse technology combined with decision analysis in the minimum living security of civil affairs departments is an inevitable trend. In this paper, the author combine these emerging technologies to build the minimum living security decision support system based on the data mining.Firstly, the paper reviews decision support systems, data warehouse, online analytical processing and data mining technology.Then referring to the provincial minimum living security information system for the study, the paper analyses the present condition of the minimum living security decision support and existing problems combined with the home office research internship results for more than a year. Finally, determine the major functional requirements and the DSS performance requirements.Secondly, according to the functional requirements the author design the minimum living security DSS, including data warehouse, online analytical processing module, data mining module and front-end interface layer. Among them, the data mining model is mainly discussed in this study. The author construct aid demand forecasting model and subsistence expenses impact indicators evaluation model to support the predefined target. In the first model, due to ARIMA algorithm with the ability to accurately extract time series regression relationship, the paper uses it to amend the wavelet neural network prediction results through the prediction error and ensure that the WNN can approximate actual benefits demand with greater precision. In the second model, in order to eliminate the influence of subjective weighting and assure the effectiveness and practicability of the weights, GA-BP is used to determine the weights of evaluation indexes. After empowering the properties, FCM algorithm is used for clustering.Finally, use the multi-layer architecture of B/S mode to complete the implementation of decision support system based on development platform, which is constructed by J2EE distributed computing technology. Through the application of the DSS, the civil affairs at all levels analysts can accomplish multi-dimensional analysis, intelligent query and deep mining with a convenient and effective access at anytime, anywhere.Ultimately, it improves the office efficiency and decision level. In addition, through the empirical study, the results show that the improved data mining algorithms ameliorate the prediction accuracy and the effectiveness of clustering.