A Bayesian Network Datamining Approach for Modelling the Physical Condition of Copper Access Networks
D. Yearling, David J. Hand · Institution of Engineering and Technology eBooks · 2003
This chapter looks at how the condition of existing access networks can be modelled. This chapter has illustrated a straightforward probabilistic modelling approach which provides a clear view of the fault propagation process within the copper access network. At this current level of application it enables a better understanding of the aetiology of faults alongside a scalable means of building multiple models over wide areas of the network. This in turn allows analysts to identify weak nodes and thus better employ scarce resource in upgrading and surveying the network, and pin-pointing hot spots for contingency provision in force majeure. Furthermore, it is useful to compare models between areas, so enabling comparison of both underlying health and work practices. This would be helpful in better understanding differences in performance between areas, and addressing the issues relating to this. All of this has direct benefit within the current broadband delivery process, where customer satisfaction must be maintained in the face of possible loss of service.