The ID3 Prediction module is specialized software designed for applying AI models to both high-frequency (HF) and low-frequency (LF) data. These AI models serve various purposes, such as predicting potential wellbore issues or performing quality assurance and quality control (QA/QC) on incoming source data.

This software employs a combination of methodologies, Machine Learning algorithms, and a set of physical functions and equations to identify drilling data anomalies. The calculations are performed in real-time, and the prediction results can seamlessly integrate into the Real Time Performance module or other third-party visualization tools.

From a practical point of view, ID3 Prediction works as a backend tool that can be used as input for any WITSML stream. In particular, the system is able to collect the data, perform the calculation of the risk indicator, and then send back this result to the WITSML server. In this way, all the software connected to the WITSML server can retrieve the results of the model and visualize the information according to the needs. Another available approach is to directly connect the Real-time viewer of the client with ID3 Prediction, through a dedicated API.