Decision Tree Classification Algorithm Research And Improvement

Data mining, also known as knowledge discovery in database (KDD), is one of the most active fields in database. Data mining aims to discover many trustful, novel, useful and readable knowledge, rules or abstract information, which can be applied in decision, process control, information management, query process and so on. So the technique and application of data mining draw upon the attention of the international academic society and develop rapidly.Classification is an important part of data mining problems. Several classification models have been proposed over the years. Among those models, decision tree is a kind of model used in forecast. Decision tree models are simple and easy to understand, easily converted into rules. It also can be constructed relatively fast compare to some of other methods. Moreover, decision tree classifiers obtain similar and sometimes better accuracy when compared with some of other classification methods.This paper is a study on decision tree classification algorithms, which mainly includes two parts.In the first part, two decision tree classification algorithms, SLIQ and SPRINT, is studied, Because they are the most useful at present. We analysis the spend time in building a decision tree with SLIQ and SPRINT, both in serial and in parallel environment. Finally, we come to some constructive conclusions.In the second part, an improvement of SLIQ and SPRINT algorithm is proposed. SLIQ and SPRINT are all faced with fixed training set. We combine incremental methods with tree building algorithm and present a new algorithm, which is suitable for increasingly training set and is very efficient. The correctness of our improvement algorithm is also given in this part.