1.Topic:Detection of logging and geophysical data and reconstruction methods using unsupervised machine learning outlier detection techniques (UML-ODTs)

This study aimed to investigate the utilization of unsupervised machine learning outlier detection techniques (UML-ODTs) to identify and reconstruct outliers in oil and gas logging curves and to assess their impact on the quality of seismic synthetic records.

2. Installation

Via PyPI Such as:

pip install sklearn

pip install matplotlib

3.Instructions

Step 1:Data preprocessing(The code file name is data_process.py) Depending on the data, run this code if preprocessing is required, and if there is no null or zero value, then you don't need to run this code.

import pandas as pd

# Read data
file_path = 'C:/Users/Dell/Desktop/2/well_log_data.csv'
data = pd.read_csv(file_path)

# Removes a row that contains a specific missing value (-9999) and a value of 0
cleaned_data = data[(data != -9999) & (data != 0)].dropna()

# Save the processed file
output_path = 'C:/Users/Dell/Desktop/2/well_log_data_process.csv'
cleaned_data.to_csv(output_path, index=False)

Step 2:Choosing the UML-ODTS algorithm to detect outliers in logging curves (The code file name is DBSCAN.py, IF.py, Local_outlier_factor.py, One-class SVM.py)

import pandas as pd
from sklearn.cluster import DBSCAN
import matplotlib.pyplot as plt
import seaborn as sns

from sklearn.ensemble import IsolationForest

from sklearn.neighbors import LocalOutlierFactor

from sklearn.svm import OneClassSVM

Step 3:Reconstructing outliers in logging curves(The code file name is Reconstructed_Data.py)

from statsmodels.nonparametric.smoothers_lowess import lowess

Contribution

We welcome all forms of contribution. Check out 'CONTRIBUTING.md' to learn how to get started.

Licence

MIT

Contact information

If you have any questions, please contact me through [email protected].