摘要:本论文主要的研究内容是决策树算法的核心技术,包括决策树的构建算法和树剪枝算法,通过对前人各种决策树算法的研究,掌握决策树构建和优化的方法,形成较为全面的总结与比较,从而掌握各种算法的特点。
ification of the actual needs, through the construction and programming decision tree pruning optimization , classification prediction function. Therefore, the main research contents summarized as follows :
( 1 ) Decision Tree Algorithm to master the theory of knowledge ;
( 2 ) understand the achievements of previous studies typical decision tree algorithm and tree pruning algorithm, learn about the latest decision tree algorithm optimization program , master tree establishment and optimization process ;
( 3 ) According to previous studies algorithm shortcomings , the corresponding optimization program , and the formation algorithm ;
( 4 ) programming decision tree classifier, and conduct experiments , summarize experimental results and apply it.
The structure more summarized as follows:
The first chapter . Brief topics of the background and significance of the research status at home and abroad , as well as papers to the main content .
Chapter theoretical basis for the decision tree . Introduction from the macro decision tree technique involves the theoretical basis for the decision tree algorithm behind specific research foundation .
CHAPTER tree core technology research and performance comparison . Details of previous studies of the decision tree algorithm and pruning algorithm typical achievements , highlighting the achievements algorithms ID3 and C4.5 algorithms, including Decision Tree algorithm related concepts related to math , algorithms, ideas and achievements contribution process , advantages and disadvantages of the algorithm , and finally on the introduction of a typical decision tree induction algorithms compare achievements . Pruning algorithm focuses on the REP and other pruning algorithm, and described the decision tree pruning algorithm to summarize the typical comparison .
Chapter pruning algorithm optimization . Summary results of previous studies , and from analysis of its shortcomings , presents a more reasonable and effective improvement programs to improve the performance of decision tree algorithm .
CHAPTER decision tree classifier design and core algorithm performance comparison experiments . Using C # technology achievements and tree pruning algorithm, and using the experimental data on the core algorithms for comparison, the experimental results . And optimize the use of decision tree classification rules are applied to predict the application .
Chapter Summary and Outlook . This article summarizes the research work carried out , talk about their research experience , and the prospects for future research work .
A decision tree is a tree structure similar to a flowchart , wherein each internal node represents an attribute of the test , each branch representing a test output , each leaf node represents a class or the class distribution . Topmost node of the tree is the root node . The purpose of constructing a decision tree is to find the relationship between attributes and classes , using it to predict the class of unknown class record . Such a system is called a predictive decision tree classifier .
Tree notation refers to it from a group of no order , no inference rule instances tree representation of classification rules . Tree representation is a greedy algorithm, using a top-down recursive way to the tree nodes inside attribute values c
本论文由英语论文网提供整理,提供论文代写,英语论文代写,代写论文,代写英语论文,代写留学生论文,代写英文论文,留学生论文代写相关核心关键词搜索。