Е.В. Бурнаев "Примеры решения задач с разными условиями на функцию выигрыша"
Upcoming SlideShare
Loading in...5
×
 

Е.В. Бурнаев "Примеры решения задач с разными условиями на функцию выигрыша"

on

  • 370 views

Е.В. Бурнаев "Примеры решения задач с разными условиями на функцию выигрыша"...

Е.В. Бурнаев "Примеры решения задач с разными условиями на функцию выигрыша"
Место показа и дата: МФТИ, школа анализа данных (ШАД), 12.05.2012

Statistics

Views

Total Views
370
Views on SlideShare
370
Embed Views
0

Actions

Likes
0
Downloads
0
Comments
0

0 Embeds 0

No embeds

Accessibility

Categories

Upload Details

Uploaded via as Adobe PDF

Usage Rights

© All Rights Reserved

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Processing…
Post Comment
Edit your comment

Е.В. Бурнаев "Примеры решения задач с разными условиями на функцию выигрыша" Е.В. Бурнаев "Примеры решения задач с разными условиями на функцию выигрыша" Presentation Transcript

  • Fig. Network volume anomalies in large-scale IP networks. Eachmeasurement corresponds to the cumulative number of bytes between two consecutive SNMP readings.
  • Fig. Approximation of real OD flows (dashed lines) by the spline-based (SB) model (full lines) in 3 operational networks. XtSMLE(k) is the real volume of OD flow k. Xt(k) stands for the estimated OD flow k using the SB model. XtTGE(k) is the estimated OD flow k using the tomogravity estimation method.
  • Fig. (a) RRMSE(t) and (b) cumulative RRMSE (t) for 672 measurements in Abilene and GEANT networks
  • Fig. (a) RRMSE(t) and (b) cumulative RRMSE (t) for 672 measurements in Abilene and GEANT networks. Fig. RRMSD(t) for 1500 flows in a Tier-2 ISP network
  • Fig. QQ-plots for 2 residual processes from (a) Abilene and (b) GEANT Fig. Correct detection rate vs. false alarm rate for the OSBD method(solid line) and the PCA approach, considering a different number of k first principal components uk to model the normal subspace.
  • Fig. Typical realizations of anomaly detection/isolation functions for a Tier-2 network (a and b) and Abilene (c and d).
  • Fig. On-line volume anomaly detection and isolation in Abilene, usingthe SSB method. The time between consecutive measurements is 5 min