Improvement of ID3 Algorithm Implementation

Ming-hao LIU, Nan LIU, Wei SHANG

Abstract


ID3 algorithm is a classical algorithm in decision tree algorithm, commonly used in data mining. Owing to that spatial information data has the characteristics of large capacity and diversity, when ID3 algorithm does data mining for spatial information, in the process of generating decision tree, some nodes will appear special sample sets which the information gain of each classification attribute is 0 and values of result attribute are not unique in. Now, conventional implementation of ID3 algorithm can't ensure the generation of decision tree. Based on this, conventional implementation of ID3 algorithm is improved, and the fault tolerance of algorithm implementation is enhanced. Improved algorithm implementation can do data mining for spatial data set

Keywords


ID3 algorithm, Data mining, Spatial information


DOI
10.12783/dtcse/smce2017/12453

Refbacks

  • There are currently no refbacks.