A study on hepatitis disease diagnosis using multilayer neural network with levenberg marquardt training algorithm

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Computational Biology

Background:

• Hepatitis diagnosis relies on accurate and timely detection.
• Machine learning offers promising avenues for improving diagnostic accuracy.
• Previous studies have explored various computational methods for hepatitis diagnosis.

Purpose of the Study:

• To develop and evaluate a neural network model for hepatitis disease diagnosis.
• To assess the performance of a multilayer neural network trained with the Levenberg-Marquardt algorithm.
• To compare the model's accuracy against existing hepatitis diagnostic studies using the UCI machine learning database.

Main Methods:

• Utilized a multilayer neural network architecture.
• Employed the Levenberg-Marquardt algorithm for neural network training and weight updates.
• Validated the model using tenfold cross-validation on the UCI machine learning database for hepatitis diagnosis.

Main Results:

• Achieved a classification accuracy of 91.87% for hepatitis disease diagnosis.
• Demonstrated competitive performance compared to previous studies using the same dataset.
• The neural network model proved effective in distinguishing hepatitis cases.

Conclusions:

• The developed neural network model shows high potential for accurate hepatitis disease diagnosis.
• The Levenberg-Marquardt algorithm is a suitable training method for this diagnostic task.
• This approach offers a valuable computational tool for supporting clinical hepatitis diagnosis.