Big data classification based on improved parallel k-nearest neighbor

In response to the rapid growth of many sorts information, highway data has continued evolve in direction big terms scale, type, and structure, exhibiting characteristics multi-source heterogeneous data. The k-nearest neighbor (KNN) join received a lot interest recent years due its wide range applications. Processing KNN joins is time-consuming inefficient quadratic structure method . As number applications dealing with vast amounts develops, get more sophisticated. authors seek save money on computer resources by leveraging large threads multiprocessors. Six popular datasets are used apply evaluate sequential parallel performance technique. These compare method. When compared matching multi-core solution, final implementation saves computing resources. It been optimized utilize as little RAM possible, allowing it manage high-resolution photo without sacrificing efficiency. will use technique they presented using Spark Radoop. Our research validates supplied method’s efficacy scalability.