Decision Support System Design And Implementation

With the rapid development of information technology, the informationrevolution brings enterprises use information technology to improve their competitivenessin the huge space: information technology not only makes the enterprise obtain the neededinformation, but also, to promote enterprise information reuse, so as to create thecompetitive advantage of the enterprise. For the FMCG industry, always looking forbusiness intelligent data understanding and implementation way, to enhance theircompetitive strength. So the enterprise decision-making layer is decision-making isdirectly related to the enterprise of vital importance. More and more enterprises leadershipof business intelligence needs. For business intelligence for domestic softwaredevelopment very much, this paper adopts the ornamental columns BQ developmentplatform design" decision making operation control system", according to BQ of datawarehouse, data mart and related ETL, data cleaning, data mining combined form,extraction in enterprise ERP system enterprises the actual operation data, through thegraph form to the decision-making layer of the enterprise. Let the decision-making layerof the enterprise can better control their own business operations.For business intelligence, he is essentially data into information process, this processcan also be referred to as information supply chain, whose objective is to bring the mostprimitive operation data becomes decision-making layer can see, compare the image to agraphical form, and then introduces the business intelligence system and key technology.Design on the fast consumer goods industry decision support system. The final text of thesummary, put forward the existing problems to discuss.Decision support system implementation is a circular spiral process, is the collectionof existing data, the data were observed, improve business operations, to collect data,forming a cycle of continual improvement. Decision support system development andimplementation in general can be divided into5phases: requirement analysis, datacollection and pretreatment stage, DEMO validation phase, data interpretation/modelingstage (Interpreting) and knowledge based action stage (Action).1the requirements analysis phase:Decision support system to solve the general is a customer problems, how to identify and understand the needs of customers problem becomes decision support managementsystem implementation of the starting point and the key to success. The customer demandanalysis method can have multiple. But the key point is to determine the issue ofcommercial value, in simple terms, customers in the decision support system on theinvestment, solve what kind of problem, can give an enterprise to bring what kind of profit,decision-making layer needs is what? Management needs is what. To process what toproblem solve.2data collection/the preprocessing stage:Data collection and the difficulty of the customer’s actual situation to have biggerconcern, according to the customer’s existing operation mode, the collected data,according to different low level database do different design. If the customer has their ownstandards and integrity of data storage, collection is easy, whereas, in the design processmust pass ETL to a variety of medium a variety of specifications of the data through thecleaning, construction, integration and formatted into the required data, in order to followthe document design, the formation of the new data can be placed in the same database.3stage DEMO:According to the needs of the enterprise itself, making a decision operation controlsystem DEMO, for the DEMO in every subject, theme of each of the elements,corresponding to the control dimension one one confirmation. So in the process ofdevelopment is not due to corporate customers of temporary demand changing. Resultedin the development of the promotion of cost.4data interpretation/modeling stage:This is a decision support system to implement the core part. However, so far thedecision support system model training result and there is no absolute measure, onlythrough the DEMO confirmation to determine which of the various themes of thedevelopment of content and control dimension. In response to the identified data andproblems, using a variety of algorithms, partitioned data set validation and parametercomparison method to find out the relative merits of topic models. Topic model isestablished, should be on the topic model training results and the establishment process ofevaluation, to finalize the design content theme.5based on the knowledge of the action phase:On the decision support system of on-line operation stage, is to obtain the enterpriseoperation data, see operation data observation results, business data change, according tothe model the results of training to take corresponding strategy.