SVM-Based Models for Predicting WLAN Traffic

Abstract

A novel type of learning machine called support vector machine (SVM) has been receiving increasing interests in the areas ranging from its original application in pattern recognition to other applications such as regression estimation due to its remarkable generalization performance. In this paper, we employ the SVM to forecast traffic in WLANs. We study the issues of one-step-ahead prediction and multi-step-ahead prediction without any assumption on the statistical property of actual WLAN traffic. We also evaluate the performance of different prediction models using four real WLAN traffic traces. The simulation results will show that among these methods, SVM outperforms other prediction models in WLAN traffic forecasting for both one-step-ahead and multi-step-ahead prediction.Recently, there has been a significant change in the understanding of network traffic. A number of recent empirical studies of traffic measurements from a variety of working packet networks have convincingly demonstrated that wired and wireless network traffic are self-similar or long-range dependent (LRD) in nature [17][18]. This implies the existence of concentrated periods of low activity and high activity (i.e., burstiness) at a wide range of time scales. It is to say that LRD is an omnipresent property of several classes of network traffic. Predicting the future traffic level from past observations is an important component to affecting traffic control under self-similar traffic conditions. Effective optimal prediction remains a technical challenge [19]. With long-range dependent network traffic and their slow convergence properties, it becomes difficult to devise effective predictors that can be rigorously shown to have desirable properties.The choice of a prediction method is a tradeoff between the prediction interval, prediction error and computational cost. In the past several decades, many methods have been proposed for network traffic prediction. To deal with the This full text paper was peer reviewed at the direction of IEEE Communications Society subject matter experts for publication in the IEEE ICC 2006 proceedings.

Traffic Pattern Prediction Based Spectrum Sharing for Cognitive Radios

“…Generally, network traffic has a mix of self-similarity, Short Range and Long Range Dependencies (SRD and LRD) (e.g., (Kleinrock, 1993); (Paxon & Floyd, 1995); (Leland et al, 1994); (Jiang et al, 2001)). There exist concentrated periods of low activity and high activity (i.e., burstiness) in the network traffic (Feng, 2006). An accurate predictor needs to capture the traffic characteristics such as SRD and LRD, selfsimilarity and nonstationarity (Sadek & Khotanzad, 2004a).…”

“…(Shu et al, 1999) also proposes using fractional ARIMA (FARIMA) to capture the self-similarity of network traffic. However, FARIMA model is time-consuming (Feng, 2006). (Sadek & Khotanzad, 2004a) discusses a two-stage predictor.…”

“…The machine learning technique called support vector machine (SVM) can be applied to pattern recognition and other applications such as regression estimation. (Feng, 2006) employs the SVM to forecast the traffic in WLANs. It studies the issues of one-step-ahead prediction and multi-step-ahead prediction without any assumption on the statistical property of actual WLAN traffic.…”

Link Load Prediction using Support Vector Regression and Optimization

“…A function K (x, x") is called a kernel if it corresponds to a dot product in some feature space and if it satisfy Mercer"s condition [6].…”

“…The complexity of the network services had led to network congestion [1,2,3,6]. In the complex network services, it has become a very difficult task to allocate bandwidth to all users so that it is judiciously utilized by all [9,15,18].…”