Implementation of Novel Algorithm (SPruning Algorithm)
Srishti Taneja · IOSR Journal of Computer Engineering · 2014
Decision trees are very significant for taking any type of verdict related to any field.Today there is ample amount of data but that data is uncooked data therefore to make it cooked data, data mining is done.Data mining concept has been used to enhance the quality of data.Various techniques are there in data mining for growth in information technology.Those techniques are classification; clustering, association rules etc.For classifying the data best method is to generate a decision tree.By using decision tree best decisions can be made as it is hierarchal structure which can be easy to interpret.A novel algorithm is implemented which gives results faster and has enhanced features than conventional algorithms of data mining.In new approach some features of Univariate algorithm (i.e.CART algorithm) and some features of Multivariate algorithm (i.e.M5P algorithm) plus an enhanced feature of new algorithm is included.The new feature of algorithm does enhancement in performance as using it performance is improved.In novel approach pruning of files is done due to which specific data can be accessed.Time is saved by using this algorithm and user can perform the task quickly and efficiently.