This analysis explores stepwise regression to predict permeability using linear models. The data is cleaned and preprocessed, followed by fitting multiple models with and without log-transformed variables. Models are optimized using backward elimination, and their performance is evaluated using RMSE on both full data and a train-test split. The results highlight the effectiveness of log transformation and stepwise selection in improving prediction accuracy.

This analysis utilizes a dataset to evaluate the heterogeneity of permeability values and their distribution characteristics. Initially, permeability data is sorted and a percentage representation is calculated to assess variability. A linear regression model is then employed to derive a heterogeneity index, which quantifies the relative differences in permeability across the dataset.

This R Markdown document analyzes porosity and permeability in geological formations through the following steps: Data Importation: Loads geological data from a CSV file. Initial Exploration: Visualizes core porosity vs. phi.N to identify trends. Porosity Model: Develops a linear regression model to predict core porosity using phi.N and facies. Corrected Porosity: Calculates refined porosity values based on the model. Permeability Model: Creates a linear model to predict core permeability from corrected porosity and facies. Corrected Permeability: Obtains corrected permeability values using the permeability model. Visualization: Generates plots to visualize the relationships between core properties and depth. Facies Analysis: Creates a boxplot to analyze depth distribution across different facies.