Traffic Analysis and Forecasting of WiMAX Network by Using ANN, FTS and Data Mining Techniques

Abstract

Forecasting of network traffic plays a vital role in the performance analysis of a network and their characteristics. In wireless networks the maintaining of uniformity is a challenging role for that case forecasting of the network traffic plays a crucial role. The main objective of this thesis is to predict traffic forecasting models of a deployed WiMAX network based on the analysis of traffic database.
For forecasting of WiMAX traffic, three different methods such as ANN, FTS, and Data mining techniques are used in this thesis. For developing of forecasting model, the real-time data are collected from the deployed BSNL WiMAX network. In this ANN method, the forecasting is done daily and weekly basis. The collection of data is 180 samples in Mbps of two base stations (BS). In this method, the traffic prediction is emphasized with the help of train LM learning algorithm using FFMLP model. In FTS method the forecasting is done based on the first order time invariant of fuzzy time series and also explained each and every step for developing a time-invariant model. Here the forecasting is done weekly and monthly basis of real time data of six months real time data. The traffic data is recorded from two BS denoted as user A and user B. this method is mainly consisting of two steps one is fuzzification and second one developing of FLRs between linguistic variables. Another method for forecasting of WiMAX traffic data is data mining technique. This method is useful for analyzing of uniformity of a traffic across the network (WiMAX). The data mining method is efficiently analyzing the high volume of data. The collection of real data is one week and each day of the week data is recorded every 1.5hr. The data is in the form of time series and used MRA in this method. In this thesis is explained the detailed theoretical background of the proposed model. The proposed model can be focused with the help of block diagram.
The forecasting of WiMAX traffic using the above methods can be with the help of statistical analysis. Based on the results of each forecasting method calculated the RMSE value. The efficiency of forecasting of WiMAX traffic data analyzed based on RMSE value.