Enhancement of Profile SVM with Distribution Management Model
Classification techniques are used to assign category labels for the transactions. Global and local learning models are used in the classification process. Global characteristics of the data are analyzed in the global learning. Nonlinear Support Vector Machine (SVM) is a popular global learning technique. Global learning technique faces model selection problem. Kernal functions are used for the model selection process. Local learning method constructs models with local neighborhood information. Local learning techniques are applied for the data sets that are not evenly distributed. K-Nearest Neighbor (KNN) classifier is one of the local learning based classification technique. Data distribution and shape of the decision surface details are not required in KNN classifiers. Computational resource requirement is high for local learning models.