PDF Publication Title:
Text from PDF Page: 109
Figure 8. Global accuracy for all methods. ARIMAX has excelled over the rest and has provided the highest prediction accuracy, with a global mean accuracy over 93%. Other algorithms as random forest, ARIMA and logistic regression have provided very good results as well, but some of the more promising algorithms as GBM have given poorer results, due to its higher sensitivity to hyper-parameters tuning and the inability to tune them for each particular device. It is interesting the good behavior of the non-time-series methods (mainly random forest) considering that we have only used two predictors. It was also surprising the bad results from HMM and ensemble classifiers. In Figure 9 we show the mean prediction accuracy and its standard deviation for all methods, and for each of the 48 hours of prediction interval; the upper chart presents the mean prediction accuracy for a 48 hours period of prediction; the intermediate chart provides the standard deviation of the mean prediction accuracy, and, the lower chart presents the mean activity of SIMs (percentage of active SIMs) for same prediction interval One interesting outcome from the study has been to observe the periodic nature of the prediction accuracy. In Figure 9, we can observe that the periodic behavior of the accuracy over time is comparable for all the methods. We have added an additional chart to the figure (bottom chart) giving the percentage of active SIMs per hour (in the 48 hours prediction period), by looking at this chart we can see that the mean accuracy has an opposite dynamic to the percentage of active SIMs. This could explain the periodic behavior of the accuracy since it seems that the low activity periods are easier to predict. Doctoral Thesis: Novel applications of Machine Learning to NTAP - 107PDF 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 | RSS | AMP |