Prity Vijay

We are living in a big data world where enormous data as a flood is brimming from all around to spawn Data Ocean. These data are fascinating if handled appropriately or else it is nothing more than trash. An ordinary algorithm is not competent in dealing out this mammoth dataset, as they are programmed to work based on the instruction. At present machine learning and data mining is gaining esteem as it is consists of a wide range of robust algorithms, which is capable of dispensation big data. The main aspiration of this work is to recognize the performances hurdle of machine learning classification algorithm due to complexity added by imbalance dataset for training purpose. The main contribution of this work is to generate a hybridization pre-processing and resampling technique which will able to reduce the complexity due to an imbalance big datasets and thus enhances performances of ML classification algorithms during assembling a precise predictive model. The algorithm proposed in this book, Hybridization Preprocessing and Resampling Technique (HPRT) is an enhanced technique, designed to reduce the complexity of dataset.

Nous vivons dans un monde de donn es o d' normes quantit s de donn es, comme une inondation, affluent de partout pour donner naissance Data Ocean. Ces donn es sont fascinantes si elles sont trait es de mani re appropri e, sinon elles ne sont rien d'autre que des d chets. Un algorithme ordinaire n'est pas comp tent pour traiter ce gigantesque ensemble de donn es, car il est programm pour fonctionner sur la base de l'instruction. l'heure actuelle, l'apprentissage machine et l'exploration de donn es gagnent en estime car ils sont constitu s d'une large gamme d'algorithmes robustes, capables de fournir de grandes quantit s de donn es. La principale aspiration de ce travail est de reconna tre l'obstacle des performances de l'algorithme de classification par apprentissage machine, d la complexit ajout e par le d s quilibre de l'ensemble de donn es des fins de formation.