Logs Prediction

With the Logs Prediction sub-strip (AI / ML > Logs Prediction) you can use machine learning to train a model to:

• Create a new log dataset for a well for which you do not have log data of this type yet, but you do have the log data for other wells in your solution.
• Fill the gaps in a log dataset.

The model learns how to interpret data from different wellbores and different data types, and can also use zonation, lithology or wellbore data as input. It is trained how to use the dependencies and distribution of existing data to predict the missing data.

Using the Logs Prediction sub-strip, you first select the data to train the model, then select what to predict, and finally select a ML model for prediction and the prediction mode (the entire log, or filling gaps in the dataset).

Prerequisites to have buttons in the strip activated:

• At least one well that contains the logs that the trainer type can learn from.
• One well with the log(s) of interest missing.