PDF Publication Title:
Text from PDF Page: 157
Fig. 4. Architecture for the best deep learning network proposed for this work (Model 2) and alternative combination with GP (Model 3). More details in [15]. IV. RESULTS This section presents a discussion of results for the different models under evaluation. We focus the analysis on the impact of design decisions, mainly aimed at the architecture of the models and the length of the elementary network flows used for the training. We provide the following performance metrics: accuracy, precision, recall, F1-score and Area Under the ROC Curve (AUC). F1-score and AUC are particularly suitable considering the highly unbalanced distribution of QoE errors. For the definition of accuracy, F1, precision, recall and AUC we follow [15]. All performance metrics given in this section are obtained with a test dataset not used at any time during training. A. Analysis of results For this problem, we have 14 distinct QoE anomalies (labels) to be detected (7 are anomalies detected at the current time-step and other 7 are anomalies for the next time-step). The anomalies can occur simultaneously, being a multi-label classification problem, considering all the labels, but with a separate error probability calculation for each label. There are two possible ways to give results in this case: aggregated and one-by-one results. For one-by-one, we focus in a particular class (label) in isolation of the other labels, simplifying the problem to a binary classification task for each particular label (one by one). In the case of aggregated results, we try to give a summary result for all labels. There are different alternatives to perform the aggregation (micro, macro, samples, weighted), varying in the way the averaging process is done. Doctoral Thesis: Novel applications of Machine Learning to NTAP - 155PDF Image | Novel applications of Machine Learning to Network Traffic Analysis
PDF Search Title:
Novel applications of Machine Learning to Network Traffic AnalysisOriginal File Name Searched:
456453_1175348.pdfDIY PDF Search: Google It | Yahoo | Bing
Cruise Ship Reviews | Luxury Resort | Jet | Yacht | and Travel Tech More Info
Cruising Review Topics and Articles More Info
Software based on Filemaker for the travel industry More Info
The Burgenstock Resort: Reviews on CruisingReview website... More Info
Resort Reviews: World Class resorts... More Info
The Riffelalp Resort: Reviews on CruisingReview website... More Info
CONTACT TEL: 608-238-6001 Email: greg@cruisingreview.com (Standard Web Page)