| The Chinese Tobacco Industry makes an important role in the nation economy, and it is the important taxation source of our country. After China joined WTO, the international trend in the home market becomes more and more notable. In order to face the competition from the foreign market, tobacco companies should gradually change their management modes, such as increasing the informatization level, introducing market organization, knowing and managing the market with informational methods, to increase the integral ability. The management of market information is a mode that centers in market, which can find out current sales status of the market timely and accurately, and it is able to establish a detailed file of sale information for each customer. At the same time, it can help the manager to make corporation plans by setting up a completed analyzing system to control the market effectively.In this article, the writer will show concrete solutions for how to set up the management system of market information with the help of solid conditions of tobacco market in home market. It includes the background and in and abroad research status of the system, the demanding analysis of information collecting, the architecture plans of the system, and detailed plans of drawing samples from the customers. In the end, the writer will show the difficulties of designing the system and provide algorithms of predicting the demanding course and realization of the code.The main tasks of this thesis:First, how to decide sample customers. Before the collection plans, we should draw samples from the whole customers. So customer sampling algorithm must be guarantee that information collected people can bear the sampling tasks which assigned to them, and also sampling customers must be similar with whole customers in sales features, that is mean the final sample customers whose sale situation can inflect the present market. Therefore, sampling algorithm directly related to the information collecting work. Because of this, how to analyze and design the sampling algorithm is the first issue of this thesis.Next, how to manage the information that collected. On the one hand it should be incarnated in the process of large numbers of demanding, editing and gathering process. On the other hand, it should be incarnated in the validating process of sincerity of the data. Only getting enough and real sale data of the customers, the system will provide solid support for the decision of the company.At last, how to achieve the correct prediction of the demands of the market. Market demanding prediction carries out the analysis and mining of the data collected, and reflect the current demanding situations in the form of numbers. As the front sections of the tobacco supply chain, accurate and reasonable demanding prediction can give directional guidance to the following stocking and sale sections. |
