Web Mining is searching useful data from the World Wide Web repository which is divided into Content Mining, Usage Mining and Structure Mining in which Content Mining uses text, images, Audio and Video to extract useful information which is Unstructured. Web Mining is sub process of Data Mining which involves Anomaly detection, Classification, Clustering, Association Rule Mining, Regression and Summarization. This discovers patterns in large data sets involving many disciplines such as Artificial Intelligence, Machine Learning, Statistics and Database Systems. Machine Learning is the emerging technology to make the machines to predict values for new data inputs according to the previous data inputs trained with some Algorithms. Among all, the Classification is in supervised Learning of Machine learning where a training set of correctly predicted observation is available. In this paper three algorithms Naïve Bayes, Random Forest and Support Vector Machine used in Rapid Miner with 500 example dataset. Accuracy and Classification error of three algorithms compared and a chart is displayed which shows that Support vector machine is 97.40% accurate than other two algorithms.