Prediction Analysis on Web Traffic Data Using Time Series Modeling, RNN and Ensembling Techniques

Abstract

In the present day web traffic holds the major segment of Internet based traffic. This information can be retrieved by building number of visits for a particular page by number of callers which helps to know the popularity of the webpage. So predicting the web traffic for further can help to maintain unforeseen traffic load there by deducting the Slashdot effect and Flash crowd effects. In this paper we mainly pivot on forecasting the Wikipedia web traffic using Ensembling technique called Boosting – AdaBoostRegressor, RNN technique LSTM and Time series modelling technique ARIMA. Further achievement of best technique of the models has been examined.