Description

In big data classification, designing a highly parallelized learning algorithm is disputed. By applying parallel computation to different components of a learning model to solve the classification problem. An extreme learning machine tree (ELM-Tree) model based on the heuristics of uncertainty reduction is introduced to resolve the over-partitioning problem in the DT induction. In this, information entropy and ambiguity are used as the uncertainty measures for splitting decision tree (DT) nodes. The given thresholds are higher than given ratios of available spilts.It effectively reduces the computational time for bigdata classification. The complex training processes of these learning algorithms make it very difficult to implement their parallelization. The high computational complexity of FT, NBTree, LMT and LMTFAM+WT seriously downgrades the ability for the model trees to handlebigdata.The parallel ELM-Tree model is effective to reduce the computational load for bigdata classification. The complex training processes of these learning algorithms make it very difficult to implement their parallelization.