Research On Business Intelligence Technology And Its Application In Supermarket Retail Trade

The business intelligence is the technical method which transforms the data material into the information or the knowledge. The greatest value of business intelligence, through the data analysis and the excavation, lies in gaining the effective and accurately information rapidly, discovering the latent rule, forecasting the trend of development, changing the information into the assistance decision-making knowledge, then being shown up to the policy-maker by the suitable way, so they can make strategical plans. To use the data effectively, discovering the basis on policy-making is the content of business intelligence.The retail sales profession has brought about lots of information in the management process, that information contains rich management viewpoint and the market rule. How to use the precious information effectively or let them serve well for the enterprise, it becomes an urgent desire for many enterprises, even a reality difficuity. The ordinary retail sales profession information system can only provide the ordinary analysis data, but it cann't provide the three-dimensional, the multi-angles of view, or the infiltrational data, let alone to provide the forecastingly latent market information, however, the business intelligence exactly makes up these system in analysis insufficiency.Customers cann't obtain satisfies due to the available choice of rich multiplicity commodity when they are supermarket shopping, on the contrary, they are always satisfied with the supermarket to its commodity choice guidance. The effective guiding customers means for supermarket is the reasonable commodity merchandises placed. On the other hand, we may find the market and the supermarket carry-ing on each kind of promotion, for example, the most common way is giving a discount, moreover, is the entire discount, however, such discount is not the best choice for them, because when the consumer purchase certain some commodities, they will also buy some another whitout considering their discount. In this case, so long as one of these two kind of commodities being sold at a discount, the other would be sold well for the stimulation of the former, so the supermarket only needs to give a discount on one kind of commodity to achieve goals of promotes sales in two kind of commodity, thus may enhance the benefit of supermarket greatly. That, what is the basis of merchandises placed for the supermarket? How do they decide the discount? When we think over them, the usual procedure manifests two kinds, the first, make alysison on the shopping basket to the commodity, namely carry on the association rules analysis to discover strong connection rule, for example, we obtain a strong association rule: namely the customers have the possibility to purchase commodity B extremely when purchase commodity A. The other procedure is to calculate the relevance between each kind of commodity, according to the relevant size to the commodity, then carry on the clustering again, make sure that the big relevant size to one cluster of the commodity, but small to other different bunchs.In order to achieve the goal above, this research took the second procedure, standarded the behavior in purchasing the commodity simultaneously a clustering question, or we could purchase some kind of similar commoditysimultaneously easily. Based on the traditional data cluster analysis, which had been widely studied and developed completely. However, the data of supermarket shopping basket had its unique characteristic, compared with the ordinary data, each kind of commodity appeared in the identical shopping basket or not, which constituted a dual data, but the ordinary based on the Euclidian distance cluster method was only suitable for the continual data. Therefore, this article introduced specially the cluster method which designed for the high-dimensional binary reponse variable, not only this method could be possible to use for doing the model cluster analysis, but also had the good direct-viewing significance.In this paper, we mainly carried on following several aspect work: First of all, it summarized the business intelligence development present situation and the trend of development, elaborated the business intelligence in retail trade application present situation in the domestic and foreign supermarket, discusssed the vital role that business intelligence technology had palyed in the management decision-making and the supermarket management, it also introduced the goal of this paper research and the content structure. Next, through to the studying of the business intelligence concept introduction, the business intelligence architecture and the construction, elaborated the technology base that business intelligence technology could be able to support the superintendent accurately and the fast policy-making. The third, through to the three big prop technology research of business intelligence, it analyzed three big technologies vital role in the business intelligence architecture, such as the data warehouse, on-line analysis processing and the data mining. Finally, it summarized development and the application present situation of business intelligence in domestic and foreign, combining the annotation this article made to the modern supermarket retail trade, stated out the new pattern that applies the business intelligence in the supermarket management and the policy-making, and discussed the new effective method to apply business intelligence, proposed the possibly problem and the limitation in the practise process. It gave advantage to further the conformity supermarket retails various links, guaranted superintendent's level of decision, and enhanced the supermarket retail sales' operation efficiency, apart from that, this method has certain guiding sense to analyzes and the forecast, and also has certain reference value in other aspect application.The article proposed the view of the supermarket shopping commodity association cluster algorithm. It was to the beneficial discussion on business intelligence technology in the supermarket retail sales service application. After using the algorithm proposed above, to carry on the cluster analysis on the supermarket commodity, obtained the significance cluster result extremely. According the analysis on the result, we can draw some valuable conclusions, which help increase the supermarket sales income and the profit, in a word, it has the actual application value.This research had certain limitation inevitably, specific speaking, it manifested in following two aspects:Firstly, because the key research point lied in the discovery to the strong relevant among products, but not to research and analysis the discovered concealed factor in the purchases behavior at the same time, therefore it was difficultly to carry on the concrete analysis to the deep level reason that the consumer simultaneously purchased these relevant very strong commodity.Secondly, the quantity method proposed in this article also had certain limitation, because it had not given the cluster to the consumerat the same time. In other words, a better method should be able to achieve to the product simultaneously cluster as well as the consumer. Therefore it's instruction role also was limited.In summary, this article proposed a suitable supermarket retail sales commodity cluster method. The Important value of this research manifested in the research method, this article research method can apply in the supermarket retail sales profession database, cause the dealer to carry on the more accurate quantitative analysis and the science forecast, which is helpful in improving its market competition ability.This article only discussed the cluster algorithm on commodity in the supermarket, the business intelligence technology also has many aspects in the supermarket application, for example, supermarket stock commodity analysis, supermarket commodity purchase analysis, supermarket customer analysis (membership card analysis) and so on. Therefore, there some many questions being worth further studying. I will do the further research work in the following aspects:First, the commodity correlational dependence proposed above is suitable merely for the dual data. That, regarding other type data and the complex data, how to calculate the association between them, this is also a question which myself will further study.Second, using the business intelligence technology, to carry on the association analysis and the cluster analysis in the supermarket commodity, is merely an application aspect in the supermarket retail sales service. I will study further to how to widely utilize this technology in other aspects, especially in supermarket retail sales service, for example, supermarket stock commodity analysis supermarket commodity purchase analysis, supermarket customer analysis (membership card analysis) and so on.