الفهرس | Only 14 pages are availabe for public view |
Abstract K-nearest Neighbors (KNN) the state of art algorithm, classified as a member of the powerfull algorithms that used in classification and regression predictions, KNN is one of the most generally utilized part of software engineering these days. It is utilized by numerous enterprises for robotizing undertakings and doing complex information examination. However, KNN has a weakness points that made this popular classifier to be considered as a lazy classifier that didn’t use it’s training set in computing or learning rather than storing or memorizing. Which means that prediction stage will be very costly in resources and time regarding large data-sets. For the KNN popularity and wide range of use, many contributions that targeting increasing the classifications efficiency worked on (KNN) and achieved milestones on enhancing the use of KNN methods. Here we present a new Algorithm (KMKNN) that will enhance the classification methods of (KNN) in terms of improves the performance of KNN in terms of prediction performance and time efficiency with the help of the unsupervised approaches clustering algorithms, that targeting the noticble efficiency results in which will save wastage of resources and time consuming during predication stage. |