The NetPredict4 project proposes the extension of the work started in the previous projects in mainly two directions. On the one hand it will perform exhaustive evaluation of the techniques developed in multiple datasets to have a comprehensive study of performance using both numerical and categorical data features. On the other hand, the time dimension will be introduced, in order to compare the characteristics of experiments done under the same conditions at different points in time and to detect performance anomalies in almost-real time in operational environments.

The final goal of Netpredict4 is to take the automated tool for detection and explanation of anomalies that is being developed in this series of projects closer to an early alarm system. The models developed will be exhaustively evaluated, and the results will be collected in an academic paper to be submitted to a high-level international conference.

e NetPredict4 project proposes the extension of the work started in the previous projects in mainly two directions. On the one hand it will perform exhaustive evaluation of the techniques developed in multiple datasets to have a comprehensive study of performance using both numerical and categorical data features. On the other hand, the time dimension will be introduced, in order to compare the characteristics of experiments done under the same conditions at different points in time and to detect performance anomalies in almost-real time in operational environments.

The final goal of Netpredict4 is to take the automated tool for detection and explanation of anomalies that is being developed in this series of projects closer to an early alarm system. The models developed will be exhaustively evaluated, and the results will be collected in an academic paper to be submitted to a high-level international conference.