Research And Development On Decision Support System Of Real Estate Based On Data Mining

The study of Real estate industrial is a highly complex system engineering that isassociated with national, collective and personal interests and influences economy. Itsprosperity and fading is a reflection of the economic development.Sheer volume of business data from both internet and offline is churned out every dayalong with the booming of China's real estate industry. This paper focuses on the analysis oflarge scale real estate data from internet and offline. The paper offers a solution to China RealEstate Information Corporation (abbr. CRIC) which faces competitive pressure of putting thedecision support system to an important role and requiring appropriate data analysis systemarchitecture.This paper discusses the enterprise's information system and proposes a real estatedecision support system based on data mining technology with given business and functionalrequirements.The first section illustrates the overall architecture of the proposed real estate decisionsupport system. The three-layer structure of the system is represented and discussed. Businessdata and Web log data is defined as the source data; data transformation layer-ETL(Extract,Transformation and Load) includes WEB log cleaning(network spiders sweeping), andbusiness data re-aggregate,this would be suit for data mining; report presentation layer isaccording to the data needs of different departments, in the form of various charts presented tousers.Secondly, the data mining models are then built with hierarchical structure. This modelanalyzes data from time dimension, geography dimension, and physical data access dimensionin terms of the URL hierarchy. Combine the Microstrategy multi-dimensional physicalstructure model to build MOLAP(Multi-dimension online analysis processing) storage mode,data mining is easy to realize in any hierarchy of time or geography dimension.Then, the Reporting System is introduced as the front tier of the real estate decisionsupport system in accordance with a combination of Microstrategy and WEB development,to build a set of complete reporting system, users or customers can login and access the reportdata, to make real-time decisions. And decisions are data based. With this real estate decision support system, we can use advanced data mining and dataanalysis tools to analyze the behaviors of network users. Based on the predictive model, itutilizes historical data to generated useful information to allocate data in storage systemaccording to their respective hotness, thereby reducing data access latency and cost. Moreover,this paper put forward a new method based on MOLAP, realizes hot spot prediction andtrending analysis, and provides decision support for network running irregular and networkstructure design.