E4.5 Final prototype for zero-touch and proactive network management

The final phase of the prototyping for zero-touch and proactive network management is the successful integration of the Feature Store with the Kubeflow Pipelines Data Acquisition module. This critical step ensures that we could efficiently acquire and process the necessary data to support and train our machine learning model. Once the data acquisition is complete, we implement a data cleaning process to eliminate irrelevant or redundant information, thereby improving the quality and reliability of our dataset.Following the data cleaning, we establish an automation process designed to enable the neural networks to dynamically adapt to the varying number of input features. This adaptability is essential for optimizing network performance and responsiveness, allowing our system to use data effectively. By automating these processes, we not only streamline our workflow but also position our network management solution to respond proactively to emerging conditions and challenges.In this deliverable we are going to describe how the data integration, processing and cleaning is done. Additionally, we will describe the automation of the neural network inputs and the exposure of the trained model using and API call.