A comprehensive survey on machine learning for networking: evolution, applications and research opportunities

In the paper

“Undoubtedly, ML has been applied to various mundane and complex problems arising in network operation and management.”

In the paper

“Whereas, passive monitoring eliminates this overhead, at the expense of additional devices that analyze the network traffic to gather relevant information.”

In the paper

“Evidently, active monitoring introduces additional overhead due to bandwidth consumption from injected traffic.”

In the paper

“Additionally, none of the ML-based AQM schemes have considered providing a fair link share among senders and have not been tested under coexisting legacy schemes in other bottleneck links in the network.”

In the paper

“In addition, MLP-NN, though applied with a variety of parameter settings, is the most popular supervised technique, while Q-learning dominates the choice of RL-based approaches.”

In the paper

“Although the ML-based resource management schemes studied in this paper differ in terms of the feature sets, they either predict one or more QoS metrics of interest or generate an acceptance/rejection decision for an incoming request, based on a QoS estimation.”

In the paper

“While both normal and fault data are easily available for a test or simulated network, only normal data with infrequent faults are routinely available for a production network.”

In the paper

“Nonetheless, NN-based supervised approaches exhibit steady performance in terms of both accuracy and convergence.”

In the paper

“The focus of the majority of approaches is on wireless networks, where resource contention is more profound than wired networks.”

In the paper

“However, unsupervised techniques can take longer time to converge than supervised approaches, potentially missing any fault occurring before the convergence.”

In the paper

“it can detect 30 anomaly records with 17 FPs, and 26 anomaly records with 1 FP”

What the sources say

• Not checked yetcites [7] Ahmed T, Coates M, Lakhina A. Multivariate online anomaly detection using kernel recursive least squares.IEEE; 2007. pp. 625–33. source

In the paper

“On the other hand, for the flow-count time series, KOAD outperforms PCA and OCNM in terms of detection rate but at the cost of a higher FP.”

What the sources say

• Not checked yetcites [7] Ahmed T, Coates M, Lakhina A. Multivariate online anomaly detection using kernel recursive least squares.IEEE; 2007. pp. 625–33. source

In the paper

“For a dataset with known attacks, both RBF-SVM with limited and all features return a TP above 98% for the malware traces, while TP of RBF-SVM is 86.2% for normal traces.”

What the sources say

• Not checked yetcites [64] Boero L, Marchese M, Zappatore S. Support vector machine meets software defined networking in ids domain. In: Proceedings of the 29th International Teletraffic Congress (ITC), vol. 3. New York: IEEE: 2017. p. 25–30.

In the paper

“In packet time series, OCNM flags 26 out of 34 anomalies but generates 14 FPs”

What the sources say

• Not checked yetcites [7] Ahmed T, Coates M, Lakhina A. Multivariate online anomaly detection using kernel recursive least squares.IEEE; 2007. pp. 625–33. source

In the paper

“A quick glance at Tables 22, 23, and 24 will reveal that the KDD'99 dataset is the dataset of choice in most anomaly-based intrusion detection literature, where some have also employed the improved version of the dataset, NSL-KDD [438] released in 2009.”

What the sources say

• Not checked yetcites [438] Tavallaee M, Bagheri E, Lu W, Ghorbani AA. A detailed analysis of the kdd cup 99 data set.IEEE; 2009, pp. 1–6. source

In the paper

“Jiang et al. [220] challenge this notion by showcasing that the size of the cluster is not sufficient to detect anomalies and has to be coupled with the distance of the cluster from other clusters, to increase accuracy of detection.”

What the sources say

• Not checked yetcites [220] Jiang S, Song X, Wang H, Han JJ, Li QH. A clustering-based method for unsupervised intrusion detections. Pattern Recog Lett. 2006; 27(7):802–10. source

In the paper

“They evaluate their work on the KDD dataset, and observe that SOM-based anomaly detection achieves an average DR (ADR) of 89% with FP in the range of , for each individual protocol and network service; e.g.ICMP, HTTP, FTP, and Mail.”

What the sources say

• Not checked yetcites [232] Kayacik HG, Zincir-Heywood AN, Heywood MI. On the capability of an SOM-based intrusion detection system. In: Proceedings of the International Joint Conference on Neural Networks. New York: IEEE: 2003. p. 1808–13.

In the paper

“For instance, Nguyen and Armitage [339], though impactful, is now dated and only addresses traffic classification in networking.”

What the sources say

• Not checked yetcites [339] Nguyen TTT, Armitage G. A survey of techniques for internet traffic classification using machine learning. IEEE Commun Surv Tutor. 2008b; 10(4):56–76. source

In the paper

“The main obstacles include what data can be collected from and what control actions can be exercised on legacy network devices.”

In the paper

“Generally, ML is ideal for inferring solutions to problems that have a large representative dataset.”

In the paper

“The key criteria used in the selection is a combination of the year of publication, citation count and merit.”

In the paper

“Despite these advances, network operations and management still remains cumbersome, and network faults are prevalent primarily due to human error [291].”

What the sources say

• Not checked yetcites [291] Mahmoud Q. Cognitive Networks: Towards Self-Aware Networks.Wiley-Interscience; 2007. source

In the paper

“On the other hand, Klaine et al. [246], Bkassiny et al. [61] and Alsheikh et al. [18], though comprehensive in their coverage of ML techniques in networking, are specialized to specific network technology i.e. cellular network, CRN and WSN, respectively.”

What the sources say

• Not checked yetcites [246] Klaine PV, Imran MA, Onireti O, Souza RD. A survey of machine learning techniques applied to self organizing cellular networks. IEEE Commun Surv Tutor. 2017; PP(99):1.
• Not checked yetcites [61] Bkassiny M, Li Y, Jayaweera SK. A survey on machine-learning techniques in cognitive radios. IEEE Commun Surv Tutor. 2013; 15(3):1136–59. source
• Not checked yetcites [18] Alsheikh MA, Lin S, Niyato D, Tan HP. Machine learning in wireless sensor networks: Algorithms, strategies, and applications. IEEE Commun Surv Tutor. 2014; 16(4):1996–2018. source

In the paper

“Second, the network is continually evolving and the dynamics inhibit the application of a fixed set of patterns that aid in network operation and management.”

In the paper

“It is almost impossible to manually keep up with network administration, due to the continuous growth in the number of applications running in the network and the kinds of devices connected to the network.”

In the paper

“The cognition from ML can be used to aid in the automation of network operation and management tasks.”

In the paper

“The imbalance in the number of training data across classes, is a violation of the assumptions maintained by many ML techniques, that is, the data is independent and identically distributed.”

In the paper

“It is important to note that there is a strong relationship between the training data, problem and the learning paradigm.”

In the paper

“Though there are different categories of problems that enjoy the benefits of ML, there is a generic approach to building ML-based solutions.”

In the paper

“Networking problems can be formulated as one of these problems that can leverage ML.”

In the paper

“Inadvertently, the ground truth drives the accuracy of ML models.”

In the paper

“These models are good at inferring discriminators for binary, textual and structural content [286].”

What the sources say

• Not checked yetcites [286] Ma J, Levchenko K, Kreibich C, Savage S, Voelker GM. Unexpected means of protocol inference. In: Proceedings of the 6th ACM SIGCOMM conference on Internet measurement.2006. p. 313–26. source

In the paper

“Hence, RL is best suited for making cognitive choices, such as decision making, planning and scheduling [441].”

What the sources say

• Not checked yetcites [441] Tesauro G. Reinforcement learning in autonomic computing: A manifesto and case studies. IEEE Internet Comput. 2007; 11(1):22–30. source

In the paper

“For instance, in traffic classification, establishing ground truth for application classes in the training dataset can be achieved using application signature pattern matching [140].”

What the sources say

• Not checked yetcites [140] Este A, Gringoli F, Salgarelli L. Support vector machines for tcp traffic classification. Comput Netw. 2009; 53(14):2476–90. source

In the paper

“Irrelevant features increase computational overhead with marginal to no gain in accuracy, while redundant features promote over-fitting.”

In the paper

“Average packet size is fairly constant in nature [492] and packet inter-arrival times are a good discriminator for bulk data transfer (e.g.FTP) and streaming applications [390].”

What the sources say

• Not checked yetcites [492] Zander S, Nguyen T, Armitage G. Automated traffic classification and application identification using machine learning.IEEE; 2005, pp. 250–7. source

In the paper

“Therefore, typically there is a need to combat class imbalance by applying under-, over-, joint-, or ensemble-sampling techniques [267].”

What the sources say

• Not checked yetcites [267] Lemaître G, Nogueira F, Aridas CK. Imbalanced-learn: A python toolbox to tackle the curse of imbalanced datasets in machine learning. J Mach Learn Res. 2017; 18(17):1–5.

In the paper

“Unlike, generative and discriminative approaches that are myopic in nature, RL may sacrifice immediate gains for longterm rewards.”

In the paper

“However, these application signatures must be kept up-to-date and adapted to the application dynamics [176].”

What the sources say

• Not checked yetcites [176] Haffner P, Sen S, Spatscheck O, Wang D. Acas: automated construction of application signatures. In: Proceedings of the 2005 ACM SIGCOMM workshop on Mining network data. New York: ACM: 2005. p. 197–202.

In the paper

“In addition, both validation and testing datasets should be independent of the training dataset and follow the same probability distribution as the training dataset.”

In the paper

“ML techniques require representative data, possibly without bias, to build an effective ML model for a given networking problem.”

In the paper

“However, the outcomes are periodically validated, since they can drift over time, known as concept drifting.”

In the paper

“Application signatures for encrypted traffic (e.g.SSH, HTTPS) extract the signature from unencrypted handshakes.”

In the paper

“In 1977, in the area of statistical learning, Dempster et al. proposed EM, a generalization of the previous iterative, unsupervised methods, such as the Baum-Welch algorithm, for learning the unknown parameters of statistical HMM models [115].”

What the sources say

• Not checked yetcites [115] Dempster AP, Laird NM, Rubin DB. Maximum likelihood from incomplete data via the em algorithm. J R Stat Soc Ser B (Method). 1977; 39(1):1–38. source

In the paper

“Towards the end of the 1970s, the second generation of DTs emerged as the Iterative Dichotomiser 3 (ID3) algorithm was released.”

In the paper

“Though, BP is widely used in training NNs, its efficiency depends on the choice of initial weights.In particular, BP has been shown to have slower speed of convergence and to fall into local optima.”

In the paper

“In addition, the Markov model [159, 296] (elaborated 50 years earlier) was leveraged to construct a process based on discrete-time state transitions and action rewards, named Markov Decision Process (MDP), which formalizes sequential decision-making problems in a fully observable, controlled environment [46].”

What the sources say

• Not checked yetcites [159] Gagniuc P. Markov Chains: From Theory to Implementation and Experimentation.Hoboken: Wiley; 2017. source
• Not checked yetcites [296] Markov AA. An example of statistical investigation in the text of eugene onegin illustrating coupling of tests in chains. In: Proceedings of the Royal Academy of Sciences of St. Petersburg, St. Petersburg, Rusia, vol 1.1913. p. 153.
• Not checked yetcites [46] Bellman R. Dynamic Programming, 1st ed. Princeton: Princeton University Press; 1957.

In the paper

“At the same time, Witten developed an RL approach to solve MDPs, inspired by animal behavior and learning theories [478], that was later referred to as Temporal Difference (TD) in Sutton's work [433,434].”

What the sources say

• Not checked yetcites [478] Witten IH. An adaptive optimal controller for discrete-time markov environments. Inf Control. 1977; 34(4):286–95. source
• Not checked yetcites [433] Sutton RS. Learning to predict by the methods of temporal differences. Mach learn. 1988; 3(1):9–44.
• Not checked yetcites [434] Sutton RS, Barto AG. A temporal-difference model of classical conditioning. In: Proceedings of the ninth annual conference of the cognitive science society. Seattle, WA: 1987. p. 355–78.

In the paper

“Although ML research was progressing slower than projected in the 1970s [397], the 1970s were marked by milestones that greatly shaped the evolution of ML, and contributed to its success in the following years.”

What the sources say

• Not checked yetcites [397] Russell S, Norvig P. Artificial Intelligence: A Modern Approach, 3rd ed. Upper Saddle River: Prentice Hall Press; 2009.

In the paper

“By the end of the 1950s, the term “Machine Learning” was coined and defined for the first time by Arthur Samuel (cf., Section 2), who also developed a checkers-playing game that is recognized as the earliest self-learning program [401].”

What the sources say

• Not checked yetcites [401] Samuel AL. Some studies in machine learning using the game of checkers. IBM J Res Dev. 1959; 3(3):210–29. source

In the paper

“The partially observable Markov decision process (POMDP) approach to finding optimal or near-optimal control strategies for partially observable stochastic environments, given a complete model of the environment, was first proposed by Cassandra et al. [25] in 1965, while the algorithm to find the optimal solution was only devised 5 years later”

What the sources say

• Not checked yetcites [25] Astrom K. Optimal control of markov processes with incomplete state information. J Mathl Anal Appl. 1965; 10(1):174–205. source

In the paper

“Over the years, global optimization methods have been proposed to replace BP, including Genetic Algorithms (GA), Simulated Annealing (SA), and Ant Colony (AC) algorithm [500].”

What the sources say

• Not checked yetcites [500] Zhu Y, Zhang G, Qiu J. Network traffic prediction based on particle swarm bp neural network. JNW. 2013; 8(11):2685–91.

In the paper

“For example, uniform weighted threshold under-sampling creates smaller balanced training sets [222].”

What the sources say

• Not checked yetcites [222] Jin Y, Duffield N, Erman J, Haffner P, Sen S, Zhang ZL. A modular machine learning system for flow-level traffic classification in large networks. ACM Trans Knowl Discov Data (TKDD). 2012; 6(1):4.

In the paper

“In the meantime, the first algorithm for training MLP-NNs with many layers -also known as Deep NN (DNN) in today's jargon-was published by Ivakhnenko and Lapa in 1965 [210].”

What the sources say

• Not checked yetcites [210] Ivakhnenko A, Lapa V, ENGINEERING PULISOE. Cybernetic Predicting Devices. Purdue University School of Electrical Engineering; 1965.

In the paper

“In the area of clustering, Steinhaus [422] was the first to propose a continuous version of the to be called k-Means algorithm [290], to partition a heterogeneous solid with a given internal mass distribution into k subsets.”

What the sources say

• Not checked yetcites [422] Steinhaus H. Sur la division des corp materiels en parties. Bull Acad Polon Sci. 1956; 1:801–4.
• Not checked yetcites [290] MacQueen J. Some methods for classification and analysis of multivariate observations. In: Proceedings of the Fifth Berkeley Symposium on Mathematical Statistics and Probability, Volume 1: Statistics, University of California Press, Berkeley, 1: 1967. p. 281–97.

In the paper

“Although CMAC was primarily designed as a function modeler for robotic controllers, it has been extensively used in RL and classification problems for its faster learning compared to MLP.”

In the paper

“In 1975, Albus proposed CMAC, a new type of NN as an alternative to MLP [11].”

What the sources say

• Not checked yetcites [11] Albus JS. A new approach to manipulator control: the cerebellar model articulation controller (cmac). J Dyn Syst Meas. Control. 1975; 97:220–7. source

In the paper

“BP is to date the most popular NN training algorithm, and comes in different variants such as Gradient Descent (GD), Conjugate Gradient (CG), One Step Secant (SS), Levenberg-Marquardt (LM), and Resilient backpropagation (Rp).”

In the paper

“These include the Backpropagation (BP) algorithm, the Cerebellar Model Articulation Controller (CMAC) NN model [11], the Expectation Maximization (EM) algorithm [115], the to-be-referred-to as Temporal Difference (TD) learning [478], and the Iterative Dichotomiser 3 (ID3) algorithm [373].”

What the sources say

• Not checked yetcites [11] Albus JS. A new approach to manipulator control: the cerebellar model articulation controller (cmac). J Dyn Syst Meas. Control. 1975; 97:220–7. source
• Not checked yetcites [115] Dempster AP, Laird NM, Rubin DB. Maximum likelihood from incomplete data via the em algorithm. J R Stat Soc Ser B (Method). 1977; 39(1):1–38. source
• Not checked yetcites [478] Witten IH. An adaptive optimal controller for discrete-time markov environments. Inf Control. 1977; 34(4):286–95. source

In the paper

“It should be noted that a classification model with a negative gradient in the ROC curve can be easily improved by flipping the predictions from the model [16] or swapping the labels of the actual instances.”

What the sources say

• Not checked yetcites [16] Alpaydin E. Introduction to Machine Learning, 3rd ed. Cambridge: MIT Press; 2014.

In the paper

“The algorithm, developed by Quinlan [373], relies on a novel concept for attribute selection based on entropy maximization.”

What the sources say

• Not checked yetcites [373] Quinlan J. Discovering rules form large collections of examples: A case study. Edingburgh: Expert Systems in the Micro Electronic Age Edinburgh Press; 1979.

In the paper

“By the end of the 1960s, Minsky and Papertkey's Perceptrons book [315] drew the limitations of perceptrons-based NN through mathematical analysis, marking a historical turn in AI and ML in particular, and significantly reducing the research interest for this area over the next several years [397].”

What the sources say

• Not checked yetcites [315] Minsky M, Papert S. Perceptrons: An Introduction to Computational Geometry. Cambridge: MIT Press; 1972.
• Not checked yetcites [397] Russell S, Norvig P. Artificial Intelligence: A Modern Approach, 3rd ed. Upper Saddle River: Prentice Hall Press; 2009.

In the paper

“This algorithm marked the commencement of the Deep Learning (DL) discipline, though the term only started to be used in the 1980s in the general context of ML, and in the year 2000 in the specific context of NNs [9].”

What the sources say

• Not checked yetcites [9] Aizenberg IN, Aizenberg NN, Vandewalle JP. Multi-Valued and Universal Binary Neurons: Theory, Learning and Applications. Norwell: Kluwer Academic Publishers; 2000. source

In the paper

“Werbos's application of BP -originally a control theory algorithm from the 1960s [80,81,233]-to train NNs [472] resurrected the research in the area.”

What the sources say

• Not checked yetcites [472] Werbos P. Beyond regression: New tools for prediction and analysis in the behavioral sciences. PhD thesis: Harvard University; 1975.

In the paper

“The Baum-Welch algorithm [41] was proposed in the mi-1960s to learn those conditional probabilities.”

What the sources say

• Not checked yetcites [41] Baum LE, Petrie T. Statistical inference for probabilistic functions of finite state markov chains. Ann Math Statist. 1966; 37(6):1554–63. source

In the paper

“ML research continued to flourish in the 1960s, giving rise to a novel statistical class of the Markov model, named Hidden Markov Model (HMM) [426].”

What the sources say

• Not checked yetcites [426] Stratonovich RL. Conditional markov processes. Theory Probab Appl. 1960; 5(2):156–78. source

In the paper

“The 1950s also witnessed the first applications of the Naïve Bayes (NB) classifier in the fields of pattern recognition [97] and information retrieval [297].”

What the sources say

• Not checked yetcites [97] Chow CK. An optimum character recognition system using decision functions. IRE Trans Electron Comput EC. 1957; 6(4):247–54. source
• Not checked yetcites [297] Maron ME. Automatic indexing: An experimental inquiry. J ACM. 1961; 8(3):404–17. source

In the paper

“Similarly to CNN, SOM was also designed for a specific application domain; dimensionality reduction [249].”

What the sources say

• Not checked yetcites [249] Kohonen T. Self-organized formation of topologically correct feature maps. Biol Cybern. 1982; 43(1):59–69. source

In the paper

“In 1982, the first form of Recurrent Neural Network (RNN) was introduced by Hopfield.”

In the paper

“Named after the inventor, Hopfield network is an RNN where the weights connecting the neurons are bidirectional.”

In the paper

“As such, RNNs are found particularly useful in Time Series Forecasting (TSF), handwriting recognition and speech recognition.”

In the paper

“Q-learning, however, replaces the traditional state-value function of TD by an action-value function (i.e.Q-value) that estimates the utility of taking a specific action in specific states.”

In the paper

“The second half of 1980s also witnessed the increase in popularity of the BP algorithm and its successful application in training DNNs”

In the paper

“CNN is a feedforward NN specifically designed to be applied to visual imagery analysis and classification, and thus require minimal image preprocessing.”

In the paper

“TD(λ) adds a discount factor (0 ≤ λ ≤ 1) that determines to what extent estimates of previous statevalues are eligible for updating based on current errors, in the policy evaluation process.”

In the paper

“Connectivity between neurons in CNN is inspired by the organization of the animal visual cortex -modeled by Hubel in the 1960s [200,201]-, where the visual field is divided between neurons, each responding to stimuli only in its corresponding region.”

What the sources say

• Not checked yetcites [200] Hubel DH, Wiesel TN. Receptive fields of single neurones in the cat’s striate cortex. J Physiol. 1959; 148(3):574–91. source
• Not checked yetcites [201] Hubel DH, Wiesel TN. Receptive fields, binocular interaction and functional architecture in the cat’s visual cortex. J Physiol. 1962; 160(1):106–54. source

In the paper

“In the early 1980s, three new classes of NNs emerged, namely Convolutional Neural Network (CNN) [157], Self-Organizing Map (SOM) [249], and Hopfield network [195].”

What the sources say

• Not checked yetcites [157] Fukushima K. Neocognitron: A self-organizing neural network model for a mechanism of pattern recognition unaffected by shift in position. Biol Cybern. 1980; 36(4):193–202. source
• Not checked yetcites [249] Kohonen T. Self-organized formation of topologically correct feature maps. Biol Cybern. 1982; 43(1):59–69. source
• Not checked yetcites [195] Hopfield JJ. Neural networks and physical systems with emergent collective computational abilities. Proc Natl Acad Sci. 1982; 79(8):2554–8. http://www.pnas.org/content/79/8/2554.full.pdf . source

In the paper

“Two DT learning algorithms, similar to ID3 but developed independently, referred to as Classification And Regression Trees (CART) [76], were proposed to model classification and regression problems.”

What the sources say

• Not checked yetcites [76] Breiman L, Friedman J, Stone C, Olshen R. Classification and Regression Trees. The Wadsworth and Brooks-Cole statistics-probability series. New York: Taylor & Francis; 1984.

In the paper

“Originally named Harmonium by Smolensky, RBM is a variant of Boltzmann machines [2] with the restriction that there are no connections within any of the network layers, whether it is visible or hidden.”

What the sources say

• Not checked yetcites [2] Ackley DH, Hinton GE, Sejnowski TJ. A learning algorithm for boltzmann machines. Cogn Sci. 1985; 9(1):147–69. source

In the paper

“Towards the end of 1980s, TLFN -an MLP that incorporates the time dimension into the model for conducting TSF [260]-, and RBFNN -an NN with a weighted set of Radial Basis Function (RBF) kernels that can be trained in unsupervised and supervised ways [78]-joined the growing list of NN classes.”

What the sources say

• Not checked yetcites [260] Lapedes AS, Farber RM. How neural nets work.1987.
• Not checked yetcites [78] Broomhead DS, Lowe D. Radial basis functions, multi-variable functional interpolation and adaptive networks. Memorandum No. 4148. Malvern: Royal Signals and Radar Establishment; 1988.

In the paper

“Another DT algorithm, under the name of Reduced Error Pruning Tree (REPTree), was also introduced for classification.REPTree aimed at building faster and simpler tree models using information gain for splitting, along with reduced-error pruning [374].”

What the sources say

• Not checked yetcites [374] Quinlan JR. Simplifying decision trees. Int J Man-Mach Stud. 1987; 27(3):221–34. source

In the paper

“The modern definition of RNN, as a network where connections between neurons exhibit one or more than one cycle, was introduced by Jordan in 1986 [226].”

What the sources say

• Not checked yetcites [226] Jordan MI. Serial order: A parallel distributed processing approach. Tech. Rep. ICS Report 8604. San Diego: University of California; 1986.

In the paper

“As of today, Q-learning is the most well-studied and widely-used model-free RL algorithm.”

In the paper

“Distributed representation establishes a many-to-many relationship between neurons and (feature,value) pairs for improved efficiency, such that a (feature,value) input is represented by a pattern of activity across neurons as opposed to being locally represented by a single neuron.”

In the paper

“SOMs employ an unsupervised competitive learning approach, unlike traditional NNs that apply error-correction learning (such as BP with gradient descent).”

In the paper

“The 1980s witnessed a renewed interest in ML research, and in particular in NNs.”

In the paper

“ID3 is a precursor to the popular and widely used C4.5 and C5.0 algorithms.”

In the paper

“Cycles provide a structure for internal states or memory allowing RNNs to process arbitrary sequences of inputs.”

In the paper

“Indeed any of the above NNs can be employed in a DL architecture, either by implementing a larger number of hidden layers or stacking multiple simple NNs.”

In the paper

“Among these techniques, Bayesian Network (BN) arose as a Directed Acyclic Graph (DAG) representation model for the statistical models in use [352], such as NB and HMM -the latter considered as the simplest dynamic BN [107, 110]-.”

What the sources say

• Not checked yetcites [352] Pearl J. Bayesian networks: A model of self-activated memory for evidential reasoning. Irvine: University of California; 1985, pp. 329–34.
• Not checked yetcites [107] Dagum P, Galper A, Horvitz EJ. Temporal Probabilistic Reasoning: Dynamic Network Models for Forecasting. Stanford: Knowledge Systems Laboratory, Medical Computer Science, Stanford University; 1991.
• Not checked yetcites [110] Dean T, Kanazawa K. A model for reasoning about persistence and causation. Comput Intell. 1989; 5(2):142–50. source

In the paper

“RBMs are found useful in a variety of application domains such as dimensionality reduction, feature learning, and classification, as they can be trained in both supervised and unsupervised ways.”

In the paper

“This restriction allows for more efficient and simpler learning compared to traditional Boltzmann machines.”

In the paper

“Several key concepts emerged from the 1980s' connectionism movement, one of which is the concept of distributed representation [187].Introduced by Hinton in 1986, this concept supports the idea that a system should be represented by many features and that each feature may have different values.”

What the sources say

• Not checked yetcites [187] Hinton GE, McClelland JL, Rumelhart DE. Parallel distributed processing: Explorations in the microstructure of cognition, vol. 1 In: Rumelhart DE, McClelland JL, PDP Research Group C, editors. Parallel Distributed Processing: Explorations in the Microstructure of Cognition. Cambridge, MA, USA, chap Distributed Representations: MIT Press: 1986. p. 77–109.

In the paper

“Interestingly, although the use of the term “deep learning” in the ML community dates back to 1986 [111], it did not apply to NNs at that time.”

What the sources say

• Not checked yetcites [111] Dechter R. Learning while searching in constraint-satisfaction-problems. In: Proceedings of the Fifth AAAI National Conference on Artificial Intelligence, AAAI Press, AAAI’86.Palo Alto: AAAI Press: 1986. p. 178–83.

In the paper

“Towards the end of 1980s, two TD approaches were proposed for reinforcement learning: TD(λ) [433] and Qlearning [471].”

What the sources say

• Not checked yetcites [433] Sutton RS. Learning to predict by the methods of temporal differences. Mach learn. 1988; 3(1):9–44.
• Not checked yetcites [471] Watkins CJ. Models of delayed reinforcement learning. PhD thesis: Psychology Department, Cambridge University; 1989.

In the paper

“Research on regression analysis propounded the Least Absolute Selection and Shrinkage Operator (LASSO), which performs variable selection and regularization for higher prediction accuracy [445].”

What the sources say

• Not checked yetcites [445] Tibshirani R. Regression shrinkage and selection via the lasso. J R Stat Soc Ser B (Methodol). 1996; 58(1):267–288. source

In the paper

“A singleclass SVM is a binary classifier that deduces the hyperplane to differentiate between the data belonging to the class against the rest of the data, that is, one-vs-rest.”

In the paper

“SVM enables plugging different kernel functions (e.g.linear, polynomial, RBF) to find the optimal solution in higherdimensional feature spaces.”

In the paper

“One well-known variant of the M5 algorithm is M5P, which aims at building trees for regression models.”

In the paper

“Research efforts in the field generated some of the the most widely used implementations of ensemble DT as of today: Multiple Additive Regression Trees (MART) [154], extra-trees [164], and eXtreme Gradient Boosting (XGBoost) [93].”

What the sources say

• Not checked yetcites [154] Friedman JH. Greedy function approximation: A gradient boosting machine. Ann Statist. 2001; 29(5):1189–232. source
• Not checked yetcites [164] Geurts P, Ernst D, Wehenkel L. Extremely randomized trees. Mach Learn. 2006; 63(1):3–42. source
• Not checked yetcites [93] Chen T, Guestrin C. XGBoost: A Scalable Tree Boosting System. In: Proceedings of the 22Nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining. New York: ACM: 2016. p. 785–94.

In the paper

“These advances in ML facilitated the successful deployment of major use cases in the 1990s, particularly, handwriting recognition [419] and data mining [3].”

What the sources say

• Not checked yetcites [419] Srihari SN, Kuebert EJ. Integration of hand-written address interpretation technology into the united states postal service remote computer reader system. In: Proceedings of the 4th International Conference on Document Analysis and Recognition, ICDAR ’97. Washington: IEEE Computer Society: 1997. p. 892–6.
• Not checked yetcites [3] ACM Special Interest Group on KDD. KDD cup archives. 2016. http://www.kdd.org/kdd-cup . Accessed 22 Nov 2017.

In the paper

“The 21st century began with a new wave of increasing interest in SVM and ensemble learning, and in particular ensemble DT.”

In the paper

“SVM-based classifiers find a hyperplane to discriminate between categories.”

In the paper

“MART and XGBoost are respectively a commercial and open source implementation of Friedman's Gradient Boosting Decision Tree (GBDT) algorithm; an ensemble DT algorithm based on gradient boosting [154,155].”

What the sources say

• Not checked yetcites [154] Friedman JH. Greedy function approximation: A gradient boosting machine. Ann Statist. 2001; 29(5):1189–232. source
• Not checked yetcites [155] Friedman JH. Stochastic gradient boosting. Comput Stat Data Anal. 2002; 38(4):367–78. source

In the paper

“Bioinspired optimization algorithms, such as Genetic Algorithms (GA) and Particle Swarm Optimization (PSO), received increasing attention and were used to train NNs for improved performance over the traditional BP-based learning [234,319].”

What the sources say

• Not checked yetcites [234] Kennedy J, Eberhart R. Particle swarm optimization. In: Neural Networks, 1995. Proceedings., IEEE International Conference on, vol 4.1995. p. 1942–8. source
• Not checked yetcites [319] Montana DJ, Davis L. Training feedforward neural networks using genetic algorithms. In: Proceedings of the 11th International Joint Conference on Artificial Intelligence - Volume 1, Morgan Kaufmann Publishers Inc., San Francisco, CA, USA, IJCAI’89.1989. p. 762–7.

In the paper

“In addition, some variants of the original boosting algorithm were also developed, such as Adaptive Boosting (AdaBoost) [153] and Stochastic Gradient Boosting (SGBoost) [155].”

What the sources say

• Not checked yetcites [153] Freund Y, Schapire RE. Experiments with a new boosting algorithm. In: Proceedings of the Thirteenth International Conference on International Conference on Machine Learning, ICML’96. San Francisco: Morgan Kaufmann Publishers Inc.: 1996. p. 148–56.
• Not checked yetcites [155] Friedman JH. Stochastic gradient boosting. Comput Stat Data Anal. 2002; 38(4):367–78. source

In the paper

“The first instances of ensemble learning include Weighted Majority Algorithm (WMA) [279], boosting [403], bootstrap aggregating (or bagging) [75], and Random Forests (RF) [191].”

What the sources say

• Not checked yetcites [279] Littlestone N, Warmuth MK. The weighted majority algorithm. In: 30th Annual Symposium on Foundations of Computer Science.1989. p. 256–61. source
• Not checked yetcites [403] Schapire RE. The strength of weak learnability. Mach Learn. 1990; 5(2):197–227.
• Not checked yetcites [75] Breiman L. Bagging predictors. Mach Learn. 1996; 24(2):123–40.

In the paper

“SVM has been widely used primarily for classification, although a regression variant exists, known as Support Vector Regression (SVR) [70].”

What the sources say

• Not checked yetcites [70] Boser BE, Guyon IM, Vapnik VN. A training algorithm for optimal margin classifiers. In: Proceedings of the Fifth Annual Workshop on Computational Learning Theory, ACM, New York, NY, USA, COLT ’92. New York: ACM: 1992. p. 144–52.

In the paper

“A multi-class approach in SVM can be formulated as a series of single class classifiers, where the data is assigned to the class that maximizes an output function.”

In the paper

“Another well-known ML technique introduced in the 1990s was Support Vector Machines (SVM).”

In the paper

“In DT research, Quinlan published the M5 algorithm in 1992 [375] to construct tree-based multivariate linear models analogous to piecewise linear functions.”

What the sources say

• Not checked yetcites [375] Quinlan JR. Learning with continuous classes. In: Proceedings of Australian Joint Conference on Artificial Intelligence, World Scientific.1992. p. 343–8.

In the paper

“RF focused explicitly on tree models and marked the beginning of a new generation of ensemble DT.”

What the sources say

• Not checked yetcites [191] Ho TK. Random decision forests. In: Proceedings of the Third International Conference on Document Analysis and Recognition (Volume 1) - Volume 1, IEEE Computer Society, Washington, DC, USA, ICDAR ’95. Piscataway: IEEE: 1995. p. 278.

In the paper

“Unlike Q-learning, SARSA does not update the Q-value of an action based on the maximum action-value of the next state, but instead it uses the Q-value of the action chosen in the next state.”

In the paper

“Today, LSTM is widely used in speech recognition as well as natural language processing.”

In the paper

“Furthermore, from a conceptual perspective, Tom Mitchell formally defined ML: “A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, improves with experience E” [317].”

What the sources say

• Not checked yetcites [317] Mitchell TM. Machine Learning, 1st ed. New York: McGraw-Hill, Inc.; 1997.

In the paper

“In the area of RL, SARSA (State-Action-Reward-State-Action) was introduced as a more realistic, however less practical, Q-learning variation [395].”

What the sources say

• Not checked yetcites [395] Rummery GA, Niranjan M. On-line Q-learning using connectionist systems. CUED/F-INFENG/TR 166, Cambridge University Engineering Department. 1994.

In the paper

“C4.5 is to date one of the most important and widely used algorithms in ML and data mining.”

In the paper

“The latter represented a great shift to data-driven ML, and since then it has been applied in many areas (e.g., retail, finance, manufacturing, medicine, science) for processing huge amounts of data to build models with valuable use [169].”

What the sources say

• Not checked yetcites [169] Goodfellow I, Bengio Y, Courville A. Deep Learning.MIT Press; 2016. http://www.deeplearningbook.org .

In the paper

“A year later, Quinlan published C4.5 [376], that builds on and extends ID3 to address most of its practical shortcomings, including data overfitting and training with missing values.”

What the sources say

• Not checked yetcites [376] Quinlan JR. C4.5: Programs for Machine Learning. San Francisco: Morgan Kaufmann Publishers Inc.; 1993.

In the paper

“A new emerging concept called ensemble learning demonstrated that the predictive performance of a single learning model can be be improved when combined with other models [397].”

What the sources say

• Not checked yetcites [397] Russell S, Norvig P. Artificial Intelligence: A Modern Approach, 3rd ed. Upper Saddle River: Prentice Hall Press; 2009.

In the paper

“Probably one of the most important achievements in NNs was the work on Long Short-Term Memory (LSTM), an RNN capable of learning long-term dependencies for solving DL tasks that involve long input sequences [192].”

What the sources say

• Not checked yetcites [192] Hochreiter S, Schmidhuber J. Long short-term memory. Neural Comput. 1997; 9(8):1735–80. source

In the paper

“Generally, the ARIMA model is a popular approach for TSF, where autoregressive (AR) and moving average (MA) models are applied in tandem to perform auto-regression on the differenced and “stationarized” data.”

In the paper

“Existing TSF models for traffic prediction can be broadly decomposed into statistical analysis models and supervised ML models.”

In the paper

“It entails forecasting future traffic and traditionally has been addressed via time series forecasting (TSF).”

In the paper

“[141] propose a MLP-NN-based bandwidth prediction system for Grid environments and compare it to the Network Weather Service (NWS) [480] bandwidth forecasting AR models for traffic monitoring and measurement.”

What the sources say

• Not checked yetcites [141] Eswaradass A, Sun XH, Wu M. Network bandwidth predictor (nbp): A system for online network performance forecasting. In: Proceedings of 6th IEEE International Symposium on Cluster Computing and the Grid (CCGRID). IEEE: 2006. p. 4–pp. source
• Not checked yetcites [480] Wolski R. Dynamically forecasting network performance using the network weather service. Cluster Comput. 1998; 1(1):119–32. source

In the paper

“Their primary motive was to improve accuracy over traditional AR methods.This was supported by rigorous independent mathematical proofs published in the late eighties and the early nineties by Cybenko [106], Hornik [196], and Funahashi [158].”

What the sources say

• Not checked yetcites [106] Cybenko G. Approximation by superpositions of a sigmoidal function. Mathematics of Control. Signals Syst (MCSS). 1989; 2(4):303–14. source
• Not checked yetcites [196] Hornik K. Approximation capabilities of multilayer feedforward networks. Neural Netw. 1991; 4(2):251–7. source
• Not checked yetcites [158] Funahashi KI. On the approximate realization of continuous mappings by neural networks. Neural netw. 1989; 2(3):183–92. source

In the paper

“More recently, efforts have been made to reduce overhead and, or improve accuracy in traffic prediction by employing features from flows, other than traffic volume.”

In the paper

“However, no details are provided for the characteristics of the MLP employed in the study, nor the time complexity of the system compared to NWS.”

In the paper

“Statistical analysis models are typically built upon the generalized autoregressive integrated moving average (ARIMA) model, while majority of learning for traffic prediction is achieved via supervised NNs.”

In the paper

“Extra-trees stands for extremely randomized trees, an ensemble DT algorithm that builds random trees based on k randomly chosen features.”

In the paper

“Hinton's work on Deep Belief Networks (DBN), published in 2006 [188], gave a new breath and strength to research in DNNs.”

What the sources say

• Not checked yetcites [188] Hinton GE, Osindero S, Teh YW. A fast learning algorithm for deep belief nets. Neural Comput. 2006; 18(7):1527–54. source

In the paper

“Although the term Deep RL was coined recently, this concept was already discussed and applied 25 years ago [275,440].”

What the sources say

• Not checked yetcites [275] Lin LJ. Reinforcement learning for robots using neural networks. PhD thesis. Pittsburgh, PA, USA: Carnegie Mellon University; 1992. uMI Order No. GAX93-22750.
• Not checked yetcites [440] Tesauro G. Practical issues in temporal difference learning. Mach Learn. 1992; 8(3):257–77.

In the paper

“40 gigabit/s NSF TeraGrid network datasets show that the NN outperforms the NWS bandwidth forecasting models with an error rate of up to 8 and 121.9% for MLP-NN and NWS, respectively.”

What the sources say

• Not checked yetcites [141] Eswaradass A, Sun XH, Wu M. Network bandwidth predictor (nbp): A system for online network performance forecasting. In: Proceedings of 6th IEEE International Symposium on Cluster Computing and the Grid (CCGRID). IEEE: 2006. p. 4–pp. source
• Not checked yetcites [480] Wolski R. Dynamically forecasting network performance using the network weather service. Cluster Comput. 1998; 1(1):119–32. source

In the paper

“It is important to mention that the evolution in ML research has enabled improved learning capabilities which were found useful in several application domains, ranging from games, image and speech recognition, network operation and management, to self-driving cars [120].”

What the sources say

• Not checked yetcites [120] Domingos P. The Master Algorithm: How the Quest for the Ultimate Learning Machine Will Remake Our World.Basic Books; 2015.

In the paper

“Chabaa et al. [86] evaluate the performance of various BP training algorithms to adjust the weights in the MLP-NN, when applied to Internet traffic time series.They show superior performance, with respect to RMSE and RPE, of the Levenberg-Marquardt (LM) and the Resilient backpropagation (Rp) algorithms over other BP algorithms.”

What the sources say

• Not checked yetcites [86] Chabaa S, Zeroual A, Antari J. Identification and prediction of internet traffic using artificial neural networks. J Int Learn Syst Appl. 2010; 2(03):147.

In the paper

“At the same time, the popularity of DL increased significantly after the term “deep learning” was first introduced in the context of NNs in 2000 [9].”

What the sources say

• Not checked yetcites [9] Aizenberg IN, Aizenberg NN, Vandewalle JP. Multi-Valued and Universal Binary Neurons: Theory, Learning and Applications. Norwell: Kluwer Academic Publishers; 2000. source

In the paper

“The objective in TSF is to construct a regression model capable of drawing accurate correlation between future traffic volume and previously observed traffic volumes.”

In the paper

“Indeed the proposed NN-based forecasting system shows better learning ability than NWS's.”

In the paper

“This wave of popularity in deep learning has continued to this day, yielding major research advances over the years.”

In the paper

“Network traffic prediction plays a key role in network operations and management for today's increasingly complex and diverse networks.”

In the paper

“However, with the rapid growth of networks and increasing complexity of network traffic, traditional TSF models are seemingly compromised, giving rise to more advanced ML models.”

In the paper

“Cortez et al. [104] choose to use a NN ensemble (NNE) of five MLP-NN with one hidden layer each.Resilient backpropagation (Rp) training is used on SNMP traffic data collected from two different ISP networks.”

What the sources say

• Not checked yetcites [104] Cortez P, Rio M, Rocha M, Sousa P. Internet traffic forecasting using neural networks. In: Proceedings of IEEE International Joint Conference on Neural Networks (IJCNN). IEEE: 2006. p. 2635–42. source

In the paper

“However instead to computing an optimal splitpoint for each one of the k features at each node as in RF, extra-trees selects a split-point randomly for reduced computational complexity.”

In the paper

“However, the attractiveness of DNN started decreasing shortly after due to the experienced difficulty of training DNNs using BP (e.g.vanishing gradient problem), in addition to the increasing competitiveness of other ML techniques (e.g.SVM) [169].”

What the sources say

• Not checked yetcites [169] Goodfellow I, Bengio Y, Courville A. Deep Learning.MIT Press; 2016. http://www.deeplearningbook.org .

In the paper

“It also shows that in terms of time and computational complexity, NNE outperforms the other methods with an order of magnitude, and is well suited for real-time forecasting.”

What the sources say

• Not checked yetcites [104] Cortez P, Rio M, Rocha M, Sousa P. Internet traffic forecasting using neural networks. In: Proceedings of IEEE International Joint Conference on Neural Networks (IJCNN). IEEE: 2006. p. 2635–42. source

In the paper

“For example, Deep Q-Networks (DQN) -a combination of DNN and Qlearning-was proposed for mastering video games [318].”

What the sources say

• Not checked yetcites [318] Mnih V, Kavukcuoglu K, Silver D, Rusu AA, Veness J, Bellemare MG, Graves A, Riedmiller M, Fidjeland AK, Ostrovski G, Petersen S, Beattie C, Sadik A, Antonoglou I, King H, Kumaran D, Wierstra D, Legg S, Hassabis D. Human-level control through deep reinforcement learning. Nature. 2015; 518(7540):529–33. source

In the paper

“DBN introduced an efficient training strategy for deep learning models, which was further used successfully in different classes of DNNs [49,381].”

What the sources say

• Not checked yetcites [49] Bengio Y, Lamblin P, Popovici D, Larochelle H. Greedy layer-wise training of deep networks. In: Proceedings of the 19th, International Conference on Neural Information Processing Systems. NIPS’06. Cambridge: MIT Press: 2006. p. 153–60.
• Not checked yetcites [381] Ranzato M, Poultney C, Chopra S, LeCun Y. Efficient learning of sparse representations with an energy-based model. In: Proceedings of the 19th International Conference on Neural Information Processing Systems, MIT Press, Cambridge, MA, USA, NIPS’06: 2006. p. 1137–44. source

In the paper

“The development in ML -particularly, in DNNs-grew exponentially with advances in storage capacity and largescale data processing (i.e.Big Data) [169].”

What the sources say

• Not checked yetcites [169] Goodfellow I, Bengio Y, Courville A. Deep Learning.MIT Press; 2016. http://www.deeplearningbook.org .

In the paper

“One approach that is currently receiving tremendous attention is Deep RL, which incorporates deep learning models into RL for solving complex problems.”

In the paper

“In contrast to the local optimization in BP, Zhu et al. [500] propose a hybrid training algorithm that is based on global optimization, the PSO-ABC technique [98].It is an artificial bee colony (ABC) algorithm employing particle swarm optimization (PSO), an evolutionary search algorithm.”

What the sources say

• Not checked yetcites [500] Zhu Y, Zhang G, Qiu J. Network traffic prediction based on particle swarm bp neural network. JNW. 2013; 8(11):2685–91.
• Not checked yetcites [98] Chun-Feng W, Kui L, Pei-Ping S. Hybrid artificial bee colony algorithm and particle swarm search for global optimization. Math Probl Eng. 2014;2014. source

In the paper

“For instance, Li et al. [274] propose a frequency domain based method for network traffic flows, instead of just traffic volume.The focus is on predicting incoming and outgoing traffic volume on an inter-data center link dominated by elephant flows.”

What the sources say

• Not checked yetcites [274] Li Y, Liu H, Yang W, Hu D, Xu W. Inter-data-center network traffic prediction with elephant flows: IEEE; 2016b, pp. 206–13. source

In the paper

“They propose a HMM to describe the relationship between the flow count, flow volume and their temporal dynamic behavior.The Kernel Bayes Rule (KBR) and RNN with LSTM unit is used to predict future traffic volume based on current flow count.”

What the sources say

• Not checked yetcites [94] Chen Z, Wen J, Geng Y. Predicting future traffic using hidden markov models. In: Proceddings of 24th IEEE International Conference on Network Protocols (ICNP). IEEE: 2016. p. 1–6. source

In the paper

“Most importantly, SVR is able to achieve 9000 forecast per sec with 10 input samples, and shows the potential for real-time operation.”

What the sources say

• Not checked yetcites [52] Bermolen P, Rossi D. Support vector regression for link load prediction. Comput Netw. 2009; 53(2):191–201. source

In the paper

“Chen et al. [94], investigate the possibility of reducing cost of monitoring and collecting traffic volume, by inferring future traffic volume based on flow count only.”

What the sources say

• Not checked yetcites [94] Chen Z, Wen J, Geng Y. Predicting future traffic using hidden markov models. In: Proceddings of 24th IEEE International Conference on Network Protocols (ICNP). IEEE: 2016. p. 1–6. source

In the paper

“However, collecting all elephant flows at high frequencies is more expensive than byte count for traffic volume.Therefore, elephant flow information is collected at lower frequencies and interpolated to fill in the missing values, overcoming the overhead for elephant flow collection.”

In the paper

“Although NN outperforms GPR in terms of robustness to class imbalance by looking at the consistency of its TPR with varying flow size threshold, it has the lowest average TNR of below 80% in all datasets.”

What the sources say

• Not checked yetcites [365] Poupart P, Chen Z, Jaini P, Fung F, Susanto H, Geng Y, Chen L, Chen K, Jin H. Online flow size prediction for improved network routing.IEEE; 2016, pp. 1–6. source

In the paper

“At 1sec timescale, the SVR model shows a slight improvement over an AR model in terms of RMSE.A more significant 10% improvement is achieved over a MA model.”

What the sources say

• Not checked yetcites [52] Bermolen P, Rossi D. Support vector regression for link load prediction. Comput Netw. 2009; 53(2):191–201. source

In the paper

“However, the model does not seem to be fully considering the features related to the elephant flow, which may explain the inexplicable good performance of the 0-interpolation, a simple method that fills zeros for all unknown points.”

In the paper

“On the other hand, SVM has a low computational overhead and is more robust to parameter variations (e.g.time scale, number of samples) in general.However, they are usually applied to classification rather than TSF.”

In the paper

“Bermolen et al. [52] consider the prospect of applying SVR for link load forecasting.”

What the sources say

• Not checked yetcites [52] Bermolen P, Rossi D. Support vector regression for link load prediction. Comput Netw. 2009; 53(2):191–201. source

In the paper

“As the detection flow size threshold increases, less flows are tagged as elephant flows, creating class imbalance in the training dataset and leading to lower TPR.”

In the paper

“It is also noticed that considering the size of the first three packets alone leads to over 13.5% improvement, regardless of whether GPR or oBMM is used, with a very slight impact on the completion time of mice flows.”

What the sources say

• Not checked yetcites [365] Poupart P, Chen Z, Jaini P, Fung F, Susanto H, Geng Y, Chen L, Chen K, Jin H. Online flow size prediction for improved network routing.IEEE; 2016, pp. 1–6. source

In the paper

“Experiments on a 2 weeks hourly traffic measurement dataset show that PSO-ABC has higher prediction accuracy than BP, with an MSE of 0.006 and 0.011, respectively, on normalized data and has stable prediction performance.”

What the sources say

• Not checked yetcites [500] Zhu Y, Zhang G, Qiu J. Network traffic prediction based on particle swarm bp neural network. JNW. 2013; 8(11):2685–91.

In the paper

“Their models incorporate FNN, trained with BP using simple gradient descent and wavelet transform to capture both the time and frequency features of the traffic time series.”

In the paper

“Although oBMM outperforms all other approaches in one dataset with an average TPR and TNR very close to 100%, it fails miserably in the other datasets with an average TPR below 50% for one dataset.”

What the sources say

• Not checked yetcites [365] Poupart P, Chen Z, Jaini P, Fung F, Susanto H, Geng Y, Chen L, Chen K, Jin H. Online flow size prediction for improved network routing.IEEE; 2016, pp. 1–6. source

In the paper

“The RNN shows a prediction MSE of 0.3 at best, 0.05 higher than KBR and twice as much as the prediction error of an RNN fed with traffic volume instead of flow count.”

What the sources say

• Not checked yetcites [94] Chen Z, Wen J, Geng Y. Predicting future traffic using hidden markov models. In: Proceddings of 24th IEEE International Conference on Network Protocols (ICNP). IEEE: 2016. p. 1–6. source

In the paper

“Poupart et al. [365] explore the use of different ML techniques for flow size prediction and elephant flow detection.These techniques include gaussian processes regression (GPR), online bayesian moment matching (oBMM) and a (106, 60, 40, 1) MLP-NN.”

What the sources say

• Not checked yetcites [365] Poupart P, Chen Z, Jaini P, Fung F, Susanto H, Geng Y, Chen L, Chen K, Jin H. Online flow size prediction for improved network routing.IEEE; 2016, pp. 1–6. source

In the paper

“Because it might not be possible to have the appropriate counter in place, or because it might be technically difficult to conduct measurements at the required speed or granularity, non-TSF approaches can be useful.”

In the paper

“This motivates the need for ensemble learning.”

In the paper

“According to [365] a more complex MLP-NN (in terms of number of layers and neurons) might be required to achieve better accuracy in a non-TSF setting, and the performance of supervised learning techniques varies when tested on different datasets.”

What the sources say

• Not checked yetcites [365] Poupart P, Chen Z, Jaini P, Fung F, Susanto H, Geng Y, Chen L, Chen K, Jin H. Online flow size prediction for improved network routing.IEEE; 2016, pp. 1–6. source

In the paper

“Traffic classification requires the ability to accurately associate network traffic to pre-defined classes of interest.These classes of interest can be classes of applications (e.g.HTTP, FTP, WWW, DNS and P2P), applications (e.g.Skype [310], YouTube [488] and Netflix [331]), or class of service [390].”

What the sources say

• Not checked yetcites [310] Microsoft Corporation. Skype. 2017. https://www.skype.com/ . Accessed 01 Aug 2017.
• Not checked yetcites [488] YouTube LLC. YouTube. 2017. https://www.youtube.com/ . Accessed 01 Aug 2017.
• Not checked yetcites [331] Netflix Inc. Netflix. 2017. https://www.netflix.com/ . Accessed 01 Aug 2017.