A Novel Cloud based and Cyclic Approach for Supervised Learning

The Clouds provide more services to the users. In the past era the software users have to purchase the software with licence. But now a days clouds provide these softwares on pay and use basis to use it. This pay and use service brought more comfort to the users. Through this paper, we propose `A Novel Cloud based and Cyclic Approach for Supervised Learning´, in the area of Data Mining. Because the data warehouses are utilizing the virtualized Iaas cloud service. i.e., the data warehouses are stored in the clouds and utilizing services of cloud. We concentrated more on classification problem in the area of data mining, because the global business scenario is entirely changed. According to these changes the need of classification become essential in all areas. Particularly to the data, which is stored in the virtual data warehouses. In the cloud hosted data warehouses the current test data set will become as training data set after some period of time. In our proposed approach we introduced the post-mortem technique on the classification model to know the facts, how good the model is induced to classify the data set from the training data set. To provoke a high-quality and an efficient classification model, the model must go through the post-mortem operation to know the reality of the classification. The test data set must go through the pre-processing operation to make the data set pure and clean. This process must be done in routine.