Data Mining Research Of Vehicle Sales Based On Hash Quick Attribute Reduction Algorithm

In recent years, the high-speed development of computer networks andinformation technology has brought human society tremendous changes andfar-reaching impact. The data has been widely concerned to be an important strategicresource gradually. Data mining technology has also got a continued progress andimproves. Data mining is an important part of the process of knowledge discovery.The role of data mining is summed users interested in the knowledge drawn from thevast amounts of data.Data mining technology has brought people living convenienceand great economic benefits. Nevertheless, the collection of the data collected in reallife often contains noise data, and there may be incomplete, inconsistent, uncertaintyinformation data. In this low-quality data sets, data mining cannot get what we areexpected efficiency and conclusions. Data mining with based on rough set is widelyused as a data mining technique. Many outstanding experts made a huge contributionto the consummation the theory of the rough set.Hash quick attribute reduction algorithm has been put forward by Yong Liu andothers,2009. In the preliminary analysis and rough set algorithm comparison, I foundin this algorithm which the way of calculating the Positive region make a largereduce of the complexity than the original algorithm, and also make a greatly reducesof the overall time complexity of the whole algorithm. This algorithm can effectivelyremove irrelevant attributes, and has the ability to output reduction completeness.Because of the combination of data mining algorithms and the actualproduction, based on the characteristics of the vehicle sales data this thesis improvedthe Hash fast attribute reduction algorithm, the improvement has the same complexityof theory time with the original algorithm, but for the uneven distribution of attributevalues and value range complex data, the improved algorithm is superior to the original algorithm in computation time. Most of the production data are morecomplex, so the improved algorithm is more closely with the actual production data.Based on the improved algorithm, we make some attribute reduction for vehiclesales data set and get some corresponding reduction. After that this paper use thevalue reduction algorithm based on rough sets to complete the reduction and get thedecision rules table, and pick a series of decision rules. This paper does the expositionon the data analysis process, and then process the improved of the algorithm. Finallygenerate some useful decision rules. These decision rules can play an important rolein the specific sales.Improved algorithm can also be applied to the website’s try recommended.Because it is based on inconsistent data attribute reduction, so it has Stronganti-interference and strong portability.