Abstract: The present disclosure relates to a method performed by a network node configured for making automatic predictions in a telecommunication network. The method comprises obtaining a first value of a first key performance indicator (KPI) for a first network entity (NE) in the telecommunication network. The method also comprises obtaining a second value of a second KPI for a communication route between said first NE and a second NE. The method also comprises predicting, automatically and based on the obtained first and second values, that an alarm will be triggered at the second NE.

Abstract: A stream processing handling arrangement provides a number of data handling blocks, selectable for forming a work flow. The blocks comprise data processing blocks. The arrangement receives selections of data handling blocks, block parameter settings and workflow property objectives, and compares the objectives with corresponding stream processing framework capabilities. The arrangement determines a stream processing framework for the work flow, provides a number software module groups, each corresponding to a different framework and comprising software modules corresponding to the data handling blocks, and chooses a group corresponding to the determined stream processing framework. The arrangement selects, in this group, software modules corresponding to the selected blocks and constructs an application using the selected software modules.

Abstract: A method for propagating user preference information in a communications network, in which the user preference information may be available for a subset of users within the network. The method comprises generating individual user attribute vectors, based on user historical data and estimating user preference information. The method further comprises defining a community structure for the network and generating a stacked representation for users, the representation comprising the user attribute vector augmented with an aggregated vector of estimated user preferences of members of the user's community. The method further comprises learning a function relating the stacked representation to user preference, using the subset of users, and applying the learned function to users outside the subset.