AI Summary & Key Takeaways
AI InsightThe method proposed in the study has been tested on real datasets, and these tests show that the proposed method has contributed approximately 2% to the success ofthe DWKNN method.
📌 3 Key Takeaways
Abstract
Abstract: This siudy proposes new neighborhood parameter selection method for distance-weighted k—Nearest Neighbors (DWKNN) classification. According to this method, individual circular neighborhood boundaries are formed as per class. Then, these circular boundaries are respectively positioned such that their centers become test element. Membership of the test elements within classes is determined by the elements pertaining to the class which stays within the circle and constitute solely that circle. The proposed method is originally the state of circular attribute of the distance, or DWKNN The proposed solution applies the circular attribute contribution approach in the issue of neighborhood boundary selection in the DWKNN method. Because the circles constituting the neighborhood boundary in the proposed method are determined of given class siructure for the first time, and since the circles were designated according to the classification nature, the goal is maximum performance ofthe proposed model. The method proposed in the study has been tested on real datasets, and these tests show that the proposed method has contributed approximately 2% to the success ofthe DWKNN method.