Based On Decision Tree Con,-miner Data Mining Model Design And Implementation

In recent years, with the development of booming container transportation along with nations, the container transportation of our ports are booming . IN the period , more and more original data is collected and stored in computers. But the poor abi lity of managing data makes data rich and knowledge lean. It is for this reason that data mining, to discover useful knowledge from amount of data by uncommon methods, has developed very quickly. Classification is a widely used technology in data mining. There are many algorithms that have been proposed recently .decision tree algorithms is one of them .For the purpose of summary the development of the container transportation of our ports better, I designed the management system of container and the data mining model of container : CON-MINNER, leading in tutor.The model adopted the new algorithm of decision tree algorithm :CON-DM. It guaranteed the usefulness of data mining . In the process of model real izing, the language of JAVA with the thought of facing to the object is adopted, and it supports a foundation for more development.The decision tree learning algorithm is very useful in data raining technique. Firstly based on study of traditional IDS (Information Definition) , The method of improving IDS , CON-DM algorithm, is put forward by introducing the concept of correlation degree .Firstly, the CON-DM algorithm draws the Med Gen algorithm: analyzing the relativity between each condition property and classification property in the first. Then, remove the property which correlation degree less than the threshold of from the condition property set .This reduced the repetition of the subtree, lowering effectively the complicated degree of the decision tree, make born knowledge comprehended moreeasily. The next in order, the CON-DM algorithm use the composite basis to replace the information gain basis as the standard of chosing decision property. This solved the problem of that the information gain basis results in incline toward to the property which has more values in some degrees and optimize the decision tree's instruction and improve the classify cor- rectness of the decision tree .Secondly, this thesis aims at the customer requirements, design and realize the CON-MINNER data mining model on the base of the CON-DM a-Igorithm. This model primarily construct the estimate model about the earning ability of sea route according to the propertied and action of the sea routes of the earnings and the sea routes of no earnings .which serve for the business enterprise decision. That model establishes the data warehouse according to this topic first, and provided the function of usher the data warehouse. Customer can mine from different levelthrough the data cube in the data warehouse.This thesis is divided into six chapter. The background of the the thesis is introduced in chapter 1. Chapter 2 is a introduction about data mining technique. Chapter 3 is a introduction about decision tree .In chapter 4, CON-DM algorithm is introduced detailedly. Chapter 5 and chapter 6 is mainly about the CON-MINER design project and the related development, including the total system construction design, database design, data model design, business process design and the interface style design.