Big data denotes data sets that are so large in such a way that traditional data processing applications are not enough to process it. The huge amount of sensor data contains both necessary as well as unnecessary information. In order to get the necessary information from this big data, there is a need for classification technique and also we use techniques for prediction using those data sets. Here the classification is done by two different algorithms namely C5.0 and SVC (Support Vector Clustering) algorithm, where both of them are combined in proposed work to give efficient results in classification of the required data sets. C5.0 is an algorithm used to generate a decision tree which is used for classification, and for this reason it is often refer to as a statistical classifier. It performs winnowing in such a way that the decision tree becomes more accurate and removes the attributes which may be unhelpful. The SVC is a statistics clustering algorithm that does not make any presumption on the number of the clusters in the data. The performance of both classifiers was monitored and analysed. The result of the proposed work shows better classification when compared to the single use C5.0 classifier. The future weather predictions are also been calculated and saved in the form of dataset virtualization. This dataset can be retrieved by the user server and partitioned as packets. Then, it is transferred to n number of clients simultaneously by means of network interfacing unit. Privacy preservation can also be achieved by encryption and decryption while the transfer of data takes place. This process, on the whole reduces the overload of transferring entire data.
