Well-log data for machine-learning–based prediction of missing logs

This dataset contains well-log measurements used to support machine-learning–based prediction of missing geophysical logs. The data originate from five wells in the University of Kansas dataset and include continuous depth-indexed records of gamma ray (GR), sonic transit time (DT), and associated auxiliary logs.The dataset was preprocessed to remove incomplete records, followed by normalization using min–max scaling. Feature selection was performed based on Pearson correlation analysis to identify the most relevant predictors for each target variable. The data were split into training (70%) and testing (30%) subsets with a fixed random seed to ensure reproducibility.This dataset supports the development, evaluation, and reproducibility of K-nearest neighbors (KNN) regression models for predicting missing well logs in subsurface characterization studies.